deploy 2.0.0-dev-36

.vscode/resolver.code-snippets

@@ -3,8 +3,8 @@
 		"scope": "rust",
 		"prefix": "resolve",
 		"body": [
-			"impl Resolve<${1}, User> for State {",
-			"\tasync fn resolve(&self, ${1} { ${0} }: ${1}, _: User) -> anyhow::Result<${2}> {",
+			"impl Resolve<${0}> for ${1} {",
+			"\tasync fn resolve(self, _: &${0}) -> Result<Self::Response, Self::Error> {",
 			"\t\ttodo!()",
 			"\t}",
 			"}"
@@ -15,9 +15,9 @@
 		"prefix": "static",
 		"body": [
 			"fn ${1}() -> &'static ${2} {",
-			"\tstatic ${3}: OnceLock<${2}> = OnceLock::new();",
-			"\t${3}.get_or_init(|| {",
-			"\t\t${0}",
+			"\tstatic ${0}: OnceLock<${2}> = OnceLock::new();",
+			"\t${0}.get_or_init(|| {",
+			"\t\ttodo!()",
 			"\t})",
 			"}"
 		]

Cargo.toml

@@ -8,7 +8,7 @@ members = [
 ]
 
 [workspace.package]
-version = "1.19.5"
+version = "2.0.0-dev-36"
 edition = "2024"
 authors = ["mbecker20 <becker.maxh@gmail.com>"]
 license = "GPL-3.0-or-later"
@@ -26,21 +26,23 @@ environment_file = { path = "lib/environment_file" }
 environment = { path = "lib/environment" }
 interpolate = { path = "lib/interpolate" }
 formatting = { path = "lib/formatting" }
+transport = { path = "lib/transport" }
 database = { path = "lib/database" }
 response = { path = "lib/response" }
 command = { path = "lib/command" }
 config = { path = "lib/config" }
 logger = { path = "lib/logger" }
 cache = { path = "lib/cache" }
+noise = { path = "lib/noise" }
 git = { path = "lib/git" }
 
 # MOGH
 run_command = { version = "0.0.6", features = ["async_tokio"] }
-serror = { version = "0.5.1", default-features = false }
-slack = { version = "0.4.0", package = "slack_client_rs", default-features = false, features = ["rustls"] }
+serror = { version = "0.5.3", default-features = false }
+slack = { version = "1.1.0", package = "slack_client_rs", default-features = false, features = ["rustls"] }
 derive_default_builder = "0.1.8"
 derive_empty_traits = "0.1.0"
-async_timing_util = "1.0.0"
+async_timing_util = "1.1.0"
 partial_derive2 = "0.4.3"
 derive_variants = "1.0.0"
 mongo_indexed = "2.0.2"
@@ -60,63 +62,70 @@ futures-util = "0.3.31"
 arc-swap = "1.7.1"
 
 # SERVER
-tokio-tungstenite = { version = "0.27.0", features = ["rustls-tls-native-roots"] }
-axum-extra = { version = "0.10.1", features = ["typed-header"] }
+tokio-tungstenite = { version = "0.28.0", features = ["rustls-tls-native-roots"] }
+axum-extra = { version = "0.10.3", features = ["typed-header"] }
 tower-http = { version = "0.6.6", features = ["fs", "cors"] }
 axum-server = { version = "0.7.2", features = ["tls-rustls"] }
-axum = { version = "0.8.4", features = ["ws", "json", "macros"] }
+axum = { version = "0.8.6", features = ["ws", "json", "macros"] }
 
 # SER/DE
 ipnetwork = { version = "0.21.1", features = ["serde"] }
-indexmap = { version = "2.11.1", features = ["serde"] }
-serde = { version = "1.0.219", features = ["derive"] }
+indexmap = { version = "2.11.4", features = ["serde"] }
+serde = { version = "1.0.227", features = ["derive"] }
 strum = { version = "0.27.2", features = ["derive"] }
 bson = { version = "2.15.0" } # must keep in sync with mongodb version
 serde_yaml_ng = "0.10.0"
 serde_json = "1.0.145"
 serde_qs = "0.15.0"
-toml = "0.9.5"
+toml = "0.9.7"
+url = "2.5.7"
 
 # ERROR
-anyhow = "1.0.99"
-thiserror = "2.0.16"
+anyhow = "1.0.100"
+thiserror = "2.0.17"
 
 # LOGGING
-opentelemetry-otlp = { version = "0.30.0", features = ["tls-roots", "reqwest-rustls"] }
-opentelemetry_sdk = { version = "0.30.0", features = ["rt-tokio"] }
+opentelemetry-otlp = { version = "0.31.0", features = ["tls-roots", "reqwest-rustls"] }
+opentelemetry_sdk = { version = "0.31.0", features = ["rt-tokio"] }
 tracing-subscriber = { version = "0.3.20", features = ["json"] }
-opentelemetry-semantic-conventions = "0.30.0"
-tracing-opentelemetry = "0.31.0"
-opentelemetry = "0.30.0"
+opentelemetry-semantic-conventions = "0.31.0"
+tracing-opentelemetry = "0.32.0"
+opentelemetry = "0.31.0"
 tracing = "0.1.41"
 
 # CONFIG
-clap = { version = "4.5.47", features = ["derive"] }
+clap = { version = "4.5.48", features = ["derive"] }
 dotenvy = "0.15.7"
 envy = "0.4.2"
 
 # CRYPTO / AUTH
 uuid = { version = "1.18.1", features = ["v4", "fast-rng", "serde"] }
-jsonwebtoken = { version = "9.3.1", default-features = false }
+jsonwebtoken = { version = "9.3.1", default-features = false } # locked back with octorust
+rustls = { version = "0.23.32", features = ["aws-lc-rs"] }
+pem-rfc7468 = { version = "0.7.0", features = ["alloc"] }
 openidconnect = "4.0.1"
 urlencoding = "2.1.3"
 nom_pem = "4.0.0"
 bcrypt = "0.17.1"
 base64 = "0.22.1"
-rustls = "0.23.31"
+pkcs8 = "0.10.2"
+snow = "0.10.0"
 hmac = "0.12.1"
+sha1 = "0.10.6"
 sha2 = "0.10.9"
 rand = "0.9.2"
 hex = "0.4.3"
+spki = "0.7.3"
+der = "0.7.10"
 
 # SYSTEM
 portable-pty = "0.9.0"
-bollard = "0.19.2"
-sysinfo = "0.37.0"
+bollard = "0.19.3"
+sysinfo = "0.37.1"
 
 # CLOUD
 aws-config = "1.8.6"
-aws-sdk-ec2 = "1.167.0"
+aws-sdk-ec2 = "1.170.1"
 aws-credential-types = "1.2.6"
 
 ## CRON
@@ -126,14 +135,14 @@ chrono = "0.4.42"
 croner = "3.0.0"
 
 # MISC
-async-compression = { version = "0.4.30", features = ["tokio", "gzip"] }
+async-compression = { version = "0.4.32", features = ["tokio", "gzip"] }
 derive_builder = "0.20.2"
+shell-escape = "0.1.5"
 comfy-table = "7.2.1"
 typeshare = "1.0.4"
 octorust = "0.10.0"
 dashmap = "6.1.0"
 wildcard = "0.3.0"
 colored = "3.0.0"
-regex = "1.11.2"
-bytes = "1.10.1"
-shell-escape = "0.1.5"
+regex = "1.11.3"
+bytes = "1.10.1"

action/build.ts

@@ -0,0 +1,2 @@
+import { run } from "./run.ts";
+await run("build-komodo");

action/deno.json

@@ -0,0 +1,5 @@
+{
+  "imports": {
+    "@std/toml": "jsr:@std/toml"
+  }
+}

action/deploy.ts

@@ -0,0 +1,2 @@
+import { run } from "./run.ts";
+await run("deploy-komodo");

action/run.ts

@@ -0,0 +1,52 @@
+import * as TOML from "@std/toml";
+
+export const run = async (action: string) => {
+  const branch = await new Deno.Command("bash", {
+    args: ["-c", "git rev-parse --abbrev-ref HEAD"],
+  })
+    .output()
+    .then((r) => new TextDecoder("utf-8").decode(r.stdout).trim());
+
+  const cargo_toml_str = await Deno.readTextFile("Cargo.toml");
+  const prev_version = (
+    TOML.parse(cargo_toml_str) as {
+      workspace: { package: { version: string } };
+    }
+  ).workspace.package.version;
+
+  const [version, tag, count] = prev_version.split("-");
+  const next_count = Number(count) + 1;
+
+  const next_version = `${version}-${tag}-${next_count}`;
+
+  await Deno.writeTextFile(
+    "Cargo.toml",
+    cargo_toml_str.replace(
+      `version = "${prev_version}"`,
+      `version = "${next_version}"`
+    )
+  );
+
+  // Cargo check first here to make sure lock file is updated before commit.
+  const cmd = `
+cargo check
+echo ""
+
+git add --all
+git commit --all --message "deploy ${version}-${tag}-${next_count}"
+
+echo ""
+git push
+echo ""
+
+km run -y action ${action} "KOMODO_BRANCH=${branch}&KOMODO_VERSION=${version}&KOMODO_TAG=${tag}-${next_count}"
+`
+    .split("\n")
+    .map((line) => line.trim())
+    .filter((line) => line.length > 0 && !line.startsWith("//"))
+    .join(" && ");
+
+  new Deno.Command("bash", {
+    args: ["-c", cmd],
+  }).spawn();
+};

bin/binaries.Dockerfile

@@ -1,7 +1,7 @@
 ## Builds the Komodo Core, Periphery, and Util binaries
 ## for a specific architecture.
 
-FROM rust:1.89.0-bullseye AS builder
+FROM rust:1.90.0-bullseye AS builder
 RUN cargo install cargo-strip
 
 WORKDIR /builder
@@ -27,6 +27,6 @@ COPY --from=builder /builder/target/release/core /core
 COPY --from=builder /builder/target/release/periphery /periphery
 COPY --from=builder /builder/target/release/km /km
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Binaries"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/chef.binaries.Dockerfile

@@ -3,7 +3,7 @@
 
 ## Uses chef for dependency caching to help speed up back-to-back builds.
 
-FROM lukemathwalker/cargo-chef:latest-rust-1.89.0-bullseye AS chef
+FROM lukemathwalker/cargo-chef:latest-rust-1.90.0-bullseye AS chef
 WORKDIR /builder
 
 # Plan just the RECIPE to see if things have changed
@@ -31,6 +31,6 @@ COPY --from=builder /builder/target/release/core /core
 COPY --from=builder /builder/target/release/periphery /periphery
 COPY --from=builder /builder/target/release/km /km
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Binaries"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/cli/Cargo.toml

@@ -19,6 +19,7 @@ komodo_client.workspace = true
 database.workspace = true
 config.workspace = true
 logger.workspace = true
+noise.workspace = true
 # external
 futures-util.workspace = true
 comfy-table.workspace = true

bin/cli/aio.Dockerfile

@@ -1,4 +1,4 @@
-FROM rust:1.89.0-bullseye AS builder
+FROM rust:1.90.0-bullseye AS builder
 RUN cargo install cargo-strip
 
 WORKDIR /builder
@@ -20,6 +20,6 @@ ENV KOMODO_CLI_CONFIG_PATHS="/config"
 
 CMD [ "km" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo CLI"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/cli/multi-arch.Dockerfile

@@ -24,6 +24,6 @@ ENV KOMODO_CLI_CONFIG_PATHS="/config"
 
 CMD [ "km" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo CLI"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/cli/single-arch.Dockerfile

@@ -13,6 +13,6 @@ ENV KOMODO_CLI_CONFIG_PATHS="/config"
 
 CMD [ "km" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo CLI"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/cli/src/command/database.rs

@@ -2,6 +2,7 @@ use std::path::Path;
 
 use anyhow::Context;
 use colored::Colorize;
+use database::mungos::mongodb::bson::{Document, doc};
 use komodo_client::entities::{
   config::cli::args::database::DatabaseCommand, optional_string,
 };
@@ -21,6 +22,7 @@ pub async fn handle(command: &DatabaseCommand) -> anyhow::Result<()> {
     DatabaseCommand::Copy { yes, index, .. } => {
       copy(*index, *yes).await
     }
+    DatabaseCommand::V1Downgrade { yes } => v1_downgrade(*yes).await,
   }
 }
 
@@ -318,3 +320,45 @@ async fn copy(index: bool, yes: bool) -> anyhow::Result<()> {
 
   database::utils::copy(&source_db, &target_db).await
 }
+
+async fn v1_downgrade(yes: bool) -> anyhow::Result<()> {
+  let config = cli_config();
+
+  println!(
+    "\n🦎  {} Database {}  🦎",
+    "Komodo".bold(),
+    "V1 Downgrade".purple().bold()
+  );
+  println!(
+    "\n{}\n",
+    " - Downgrade the database to V1 compatible data structures."
+      .dimmed()
+  );
+  if let Some(uri) = optional_string(&config.database.uri) {
+    println!("{}: {}", " - URI".dimmed(), sanitize_uri(&uri));
+  }
+  if let Some(address) = optional_string(&config.database.address) {
+    println!("{}: {address}", " - Address".dimmed());
+  }
+  if let Some(username) = optional_string(&config.database.username) {
+    println!("{}: {username}", " - Username".dimmed());
+  }
+  println!(
+    "{}: {}\n",
+    " - Db Name".dimmed(),
+    config.database.db_name,
+  );
+
+  crate::command::wait_for_enter("run downgrade", yes)?;
+
+  let db = database::init(&config.database).await?;
+
+  db.collection::<Document>("Server")
+    .update_many(doc! {}, doc! { "$set": { "info": null } })
+    .await
+    .context("Failed to downgrade Server schema")?;
+
+  info!("V1 Downgrade complete. Ready to downgrade to komodo-core:1 ✅");
+
+  Ok(())
+}

bin/cli/src/command/execute.rs

@@ -230,6 +230,9 @@ pub async fn handle(
     Execution::GlobalAutoUpdate(data) => {
       println!("{}: {data:?}", "Data".dimmed())
     }
+    Execution::RotateAllServerKeys(data) => {
+      println!("{}: {data:?}", "Data".dimmed())
+    }
     Execution::Sleep(data) => {
       println!("{}: {data:?}", "Data".dimmed())
     }
@@ -494,6 +497,10 @@ pub async fn handle(
       .execute(request)
       .await
       .map(|u| ExecutionResult::Single(u.into())),
+    Execution::RotateAllServerKeys(request) => client
+      .execute(request)
+      .await
+      .map(|u| ExecutionResult::Single(u.into())),
     Execution::Sleep(request) => {
       let duration =
         Duration::from_millis(request.duration_ms as u64);

bin/cli/src/command/list.rs

@@ -794,7 +794,7 @@ impl PrintTable for ResourceListItem<ServerListItemInfo> {
       Cell::new(self.info.state.to_string())
         .fg(color)
         .add_attribute(Attribute::Bold),
-      Cell::new(self.info.address),
+      Cell::new(self.info.address.as_deref().unwrap_or("inbound")),
       Cell::new(self.tags.join(", ")),
     ];
     if links {

bin/cli/src/main.rs

@@ -41,6 +41,12 @@ async fn app() -> anyhow::Result<()> {
       }
       Ok(())
     }
+    args::Command::Key { command } => {
+      noise::key::command::handle(command).await
+    }
+    args::Command::Database { command } => {
+      command::database::handle(command).await
+    }
     args::Command::Container(container) => {
       command::container::handle(container).await
     }
@@ -54,9 +60,6 @@ async fn app() -> anyhow::Result<()> {
     args::Command::Update { command } => {
       command::update::handle(command).await
     }
-    args::Command::Database { command } => {
-      command::database::handle(command).await
-    }
   }
 }
 

bin/core/Cargo.toml

@@ -20,12 +20,14 @@ periphery_client.workspace = true
 environment_file.workspace = true
 interpolate.workspace = true
 formatting.workspace = true
+transport.workspace = true
 database.workspace = true
 response.workspace = true
 command.workspace = true
 config.workspace = true
 logger.workspace = true
 cache.workspace = true
+noise.workspace = true
 git.workspace = true
 # mogh
 serror = { workspace = true, features = ["axum"] }
@@ -38,7 +40,6 @@ slack.workspace = true
 svi.workspace = true
 # external
 aws-credential-types.workspace = true
-tokio-tungstenite.workspace = true
 english-to-cron.workspace = true
 openidconnect.workspace = true
 jsonwebtoken.workspace = true
@@ -70,6 +71,7 @@ chrono.workspace = true
 bcrypt.workspace = true
 base64.workspace = true
 rustls.workspace = true
+bytes.workspace = true
 tokio.workspace = true
 serde.workspace = true
 regex.workspace = true
@@ -81,3 +83,4 @@ rand.workspace = true
 hmac.workspace = true
 sha2.workspace = true
 hex.workspace = true
+url.workspace = true

bin/core/aio.Dockerfile

@@ -1,7 +1,7 @@
 ## All in one, multi stage compile + runtime Docker build for your architecture.
 
 # Build Core
-FROM rust:1.89.0-bullseye AS core-builder
+FROM rust:1.90.0-bullseye AS core-builder
 RUN cargo install cargo-strip
 
 WORKDIR /builder
@@ -58,6 +58,6 @@ ENV KOMODO_CLI_CONFIG_KEYWORDS="*config.*,*komodo.cli*.*"
 CMD [ "core" ]
 
 # Label for Ghcr
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Core"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/core/multi-arch.Dockerfile

@@ -54,6 +54,6 @@ ENV KOMODO_CLI_CONFIG_KEYWORDS="*config.*,*komodo.cli*.*"
 CMD [ "core" ]
 
 # Label for Ghcr
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Core"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/core/single-arch.Dockerfile

@@ -43,6 +43,6 @@ ENV KOMODO_CLI_CONFIG_KEYWORDS="*config.*,*komodo.cli*.*"
 CMD [ "core" ]
 
 # Label for Ghcr
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Core"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/core/src/alert/mod.rs

@@ -1,4 +1,3 @@
-use ::slack::types::Block;
 use anyhow::{Context, anyhow};
 use database::mungos::{find::find_collect, mongodb::bson::doc};
 use derive_variants::ExtractVariant;

bin/core/src/alert/slack.rs

@@ -1,3 +1,5 @@
+use ::slack::types::OwnedBlock as Block;
+
 use super::*;
 
 #[instrument(level = "debug")]
@@ -477,17 +479,20 @@ pub async fn send_alert(
     interpolator.interpolate_string(&mut url_interpolated)?;
 
     let slack = ::slack::Client::new(url_interpolated);
-    slack.send_message(text, blocks).await.map_err(|e| {
-      let replacers = interpolator
-        .secret_replacers
-        .into_iter()
-        .collect::<Vec<_>>();
-      let sanitized_error =
-        svi::replace_in_string(&format!("{e:?}"), &replacers);
-      anyhow::Error::msg(format!(
-        "Error with slack request: {sanitized_error}"
-      ))
-    })?;
+    slack
+      .send_owned_message_single(&text, blocks.as_deref())
+      .await
+      .map_err(|e| {
+        let replacers = interpolator
+          .secret_replacers
+          .into_iter()
+          .collect::<Vec<_>>();
+        let sanitized_error =
+          svi::replace_in_string(&format!("{e:?}"), &replacers);
+        anyhow::Error::msg(format!(
+          "Error with slack request: {sanitized_error}"
+        ))
+      })?;
   }
   Ok(())
 }

bin/core/src/api/execute/build.rs

@@ -249,7 +249,7 @@ impl Resolve<ExecuteArgs> for RunBuild {
         _ = cancel.cancelled() => {
           debug!("build cancelled during clone, cleaning up builder");
           update.push_error_log("build cancelled", String::from("user cancelled build during repo clone"));
-          cleanup_builder_instance(cleanup_data, &mut update)
+          cleanup_builder_instance(periphery, cleanup_data, &mut update)
             .await;
           info!("builder cleaned up");
           return handle_early_return(update, build.id, build.name, true).await
@@ -298,7 +298,7 @@ impl Resolve<ExecuteArgs> for RunBuild {
         _ = cancel.cancelled() => {
           info!("build cancelled during build, cleaning up builder");
           update.push_error_log("build cancelled", String::from("user cancelled build during docker build"));
-          cleanup_builder_instance(cleanup_data, &mut update)
+          cleanup_builder_instance(periphery, cleanup_data, &mut update)
             .await;
           return handle_early_return(update, build.id, build.name, true).await
         },
@@ -344,7 +344,8 @@ impl Resolve<ExecuteArgs> for RunBuild {
 
     // If building on temporary cloud server (AWS),
     // this will terminate the server.
-    cleanup_builder_instance(cleanup_data, &mut update).await;
+    cleanup_builder_instance(periphery, cleanup_data, &mut update)
+      .await;
 
     // Need to manually update the update before cache refresh,
     // and before broadcast with add_update.

bin/core/src/api/execute/deployment.rs

@@ -203,7 +203,8 @@ impl Resolve<ExecuteArgs> for Deploy {
     update.version = version;
     update_update(update.clone()).await?;
 
-    match periphery_client(&server)?
+    match periphery_client(&server)
+      .await?
       .request(api::container::Deploy {
         deployment,
         stop_signal: self.stop_signal,
@@ -331,8 +332,9 @@ pub async fn pull_deployment_inner(
   }
 
   let res = async {
-    let log = match periphery_client(server)?
-      .request(api::image::PullImage {
+    let log = match periphery_client(server)
+      .await?
+      .request(api::docker::PullImage {
         name: image,
         account,
         token,
@@ -414,7 +416,8 @@ impl Resolve<ExecuteArgs> for StartDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::StartContainer {
         name: deployment.name,
       })
@@ -461,7 +464,8 @@ impl Resolve<ExecuteArgs> for RestartDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::RestartContainer {
         name: deployment.name,
       })
@@ -510,7 +514,8 @@ impl Resolve<ExecuteArgs> for PauseDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::PauseContainer {
         name: deployment.name,
       })
@@ -557,7 +562,8 @@ impl Resolve<ExecuteArgs> for UnpauseDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::UnpauseContainer {
         name: deployment.name,
       })
@@ -606,7 +612,8 @@ impl Resolve<ExecuteArgs> for StopDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::StopContainer {
         name: deployment.name,
         signal: self
@@ -688,7 +695,8 @@ impl Resolve<ExecuteArgs> for DestroyDeployment {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::container::RemoveContainer {
         name: deployment.name,
         signal: self

bin/core/src/api/execute/maintenance.rs

@@ -1,12 +1,14 @@
-use std::sync::OnceLock;
+use std::{fmt::Write as _, sync::OnceLock};
 
 use anyhow::{Context, anyhow};
 use command::run_komodo_command;
 use database::mungos::{find::find_collect, mongodb::bson::doc};
 use formatting::{bold, format_serror};
+use futures::StreamExt;
 use komodo_client::{
   api::execute::{
     BackupCoreDatabase, ClearRepoCache, GlobalAutoUpdate,
+    RotateAllServerKeys,
   },
   entities::{
     deployment::DeploymentState, server::ServerState,
@@ -24,6 +26,7 @@ use crate::{
   },
   config::core_config,
   helpers::update::update_update,
+  resource::rotate_server_keys,
   state::{
     db_client, deployment_status_cache, server_status_cache,
     stack_status_cache,
@@ -317,3 +320,98 @@ impl Resolve<ExecuteArgs> for GlobalAutoUpdate {
     Ok(update)
   }
 }
+
+//
+
+/// Makes sure the method can only be called once at a time
+fn global_rotate_lock() -> &'static Mutex<()> {
+  static LOCK: OnceLock<Mutex<()>> = OnceLock::new();
+  LOCK.get_or_init(Default::default)
+}
+
+impl Resolve<ExecuteArgs> for RotateAllServerKeys {
+  async fn resolve(
+    self,
+    ExecuteArgs { user, update }: &ExecuteArgs,
+  ) -> Result<Self::Response, Self::Error> {
+    if !user.admin {
+      return Err(
+        anyhow!("This method is admin only.")
+          .status_code(StatusCode::FORBIDDEN),
+      );
+    }
+
+    let _lock = global_rotate_lock()
+      .try_lock()
+      .context("Rotate All Server Keys already in progress...")?;
+
+    let mut update = update.clone();
+
+    update_update(update.clone()).await?;
+
+    let mut servers = db_client()
+      .servers
+      .find(doc! { "config.enabled": true })
+      .await
+      .context("Failed to query servers from database")?;
+
+    let server_status_cache = server_status_cache();
+
+    let mut log = String::new();
+
+    while let Some(server) = servers.next().await {
+      let server = match server {
+        Ok(server) => server,
+        Err(e) => {
+          warn!("Failed to parse Server | {e:#}");
+          continue;
+        }
+      };
+      let Some(status) = server_status_cache.get(&server.id).await
+      else {
+        let _ = write!(
+          &mut log,
+          "\nSkipping {}: No Status ⚠️",
+          bold(&server.name)
+        );
+        continue;
+      };
+      if !matches!(status.state, ServerState::Ok) {
+        let _ = write!(
+          &mut log,
+          "\nSkipping {}: {} ⚠️",
+          bold(&server.name),
+          status.state
+        );
+        continue;
+      }
+      match rotate_server_keys(&server).await {
+        Ok(_) => {
+          let _ = write!(
+            &mut log,
+            "\nRotated keys for {} ✅",
+            bold(&server.name)
+          );
+        }
+        Err(e) => {
+          update.push_error_log(
+            "Key Rotation Failure",
+            format_serror(
+              &e.context(format!(
+                "Failed to rotate {} keys",
+                bold(&server.name)
+              ))
+              .into(),
+            ),
+          );
+        }
+      }
+    }
+
+    update.push_simple_log("Rotate Server Keys", log);
+    update.finalize();
+    update_update(update.clone()).await?;
+
+    Ok(update)
+  }
+}

bin/core/src/api/execute/mod.rs

@@ -149,6 +149,7 @@ pub enum ExecuteRequest {
   ClearRepoCache(ClearRepoCache),
   BackupCoreDatabase(BackupCoreDatabase),
   GlobalAutoUpdate(GlobalAutoUpdate),
+  RotateAllServerKeys(RotateAllServerKeys),
 }
 
 pub fn router() -> Router {

bin/core/src/api/execute/repo.rs

@@ -105,7 +105,7 @@ impl Resolve<ExecuteArgs> for CloneRepo {
     let server =
       resource::get::<Server>(&repo.config.server_id).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     // interpolate variables / secrets, returning the sanitizing replacers to send to
     // periphery so it may sanitize the final command for safe logging (avoids exposing secret values)
@@ -220,7 +220,7 @@ impl Resolve<ExecuteArgs> for PullRepo {
     let server =
       resource::get::<Server>(&repo.config.server_id).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     // interpolate variables / secrets, returning the sanitizing replacers to send to
     // periphery so it may sanitize the final command for safe logging (avoids exposing secret values)
@@ -463,7 +463,7 @@ impl Resolve<ExecuteArgs> for BuildRepo {
       _ = cancel.cancelled() => {
         debug!("build cancelled during clone, cleaning up builder");
         update.push_error_log("build cancelled", String::from("user cancelled build during repo clone"));
-        cleanup_builder_instance(cleanup_data, &mut update)
+        cleanup_builder_instance(periphery, cleanup_data, &mut update)
           .await;
         info!("builder cleaned up");
         return handle_builder_early_return(update, repo.id, repo.name, true).await
@@ -510,7 +510,8 @@ impl Resolve<ExecuteArgs> for BuildRepo {
 
     // If building on temporary cloud server (AWS),
     // this will terminate the server.
-    cleanup_builder_instance(cleanup_data, &mut update).await;
+    cleanup_builder_instance(periphery, cleanup_data, &mut update)
+      .await;
 
     // Need to manually update the update before cache refresh,
     // and before broadcast with add_update.

bin/core/src/api/execute/server.rs

@@ -50,7 +50,7 @@ impl Resolve<ExecuteArgs> for StartContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::StartContainer {
@@ -104,7 +104,7 @@ impl Resolve<ExecuteArgs> for RestartContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::RestartContainer {
@@ -160,7 +160,7 @@ impl Resolve<ExecuteArgs> for PauseContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::PauseContainer {
@@ -214,7 +214,7 @@ impl Resolve<ExecuteArgs> for UnpauseContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::UnpauseContainer {
@@ -270,7 +270,7 @@ impl Resolve<ExecuteArgs> for StopContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::StopContainer {
@@ -332,7 +332,7 @@ impl Resolve<ExecuteArgs> for DestroyContainer {
     // Send update after setting action state, this way frontend gets correct state.
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::RemoveContainer {
@@ -387,7 +387,8 @@ impl Resolve<ExecuteArgs> for StartAllContainers {
 
     update_update(update.clone()).await?;
 
-    let logs = periphery_client(&server)?
+    let logs = periphery_client(&server)
+      .await?
       .request(api::container::StartAllContainers {})
       .await
       .context("failed to start all containers on host")?;
@@ -437,7 +438,8 @@ impl Resolve<ExecuteArgs> for RestartAllContainers {
 
     update_update(update.clone()).await?;
 
-    let logs = periphery_client(&server)?
+    let logs = periphery_client(&server)
+      .await?
       .request(api::container::RestartAllContainers {})
       .await
       .context("failed to restart all containers on host")?;
@@ -489,7 +491,8 @@ impl Resolve<ExecuteArgs> for PauseAllContainers {
 
     update_update(update.clone()).await?;
 
-    let logs = periphery_client(&server)?
+    let logs = periphery_client(&server)
+      .await?
       .request(api::container::PauseAllContainers {})
       .await
       .context("failed to pause all containers on host")?;
@@ -539,7 +542,8 @@ impl Resolve<ExecuteArgs> for UnpauseAllContainers {
 
     update_update(update.clone()).await?;
 
-    let logs = periphery_client(&server)?
+    let logs = periphery_client(&server)
+      .await?
       .request(api::container::UnpauseAllContainers {})
       .await
       .context("failed to unpause all containers on host")?;
@@ -591,7 +595,8 @@ impl Resolve<ExecuteArgs> for StopAllContainers {
 
     update_update(update.clone()).await?;
 
-    let logs = periphery_client(&server)?
+    let logs = periphery_client(&server)
+      .await?
       .request(api::container::StopAllContainers {})
       .await
       .context("failed to stop all containers on host")?;
@@ -641,7 +646,7 @@ impl Resolve<ExecuteArgs> for PruneContainers {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
       .request(api::container::PruneContainers {})
@@ -686,10 +691,10 @@ impl Resolve<ExecuteArgs> for DeleteNetwork {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
-      .request(api::network::DeleteNetwork {
+      .request(api::docker::DeleteNetwork {
         name: self.name.clone(),
       })
       .await
@@ -748,10 +753,10 @@ impl Resolve<ExecuteArgs> for PruneNetworks {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
-      .request(api::network::PruneNetworks {})
+      .request(api::docker::PruneNetworks {})
       .await
       .context(format!(
         "failed to prune networks on server {}",
@@ -791,10 +796,10 @@ impl Resolve<ExecuteArgs> for DeleteImage {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
-      .request(api::image::DeleteImage {
+      .request(api::docker::DeleteImage {
         name: self.name.clone(),
       })
       .await
@@ -850,10 +855,10 @@ impl Resolve<ExecuteArgs> for PruneImages {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log =
-      match periphery.request(api::image::PruneImages {}).await {
+      match periphery.request(api::docker::PruneImages {}).await {
         Ok(log) => log,
         Err(e) => Log::error(
           "prune images",
@@ -891,10 +896,10 @@ impl Resolve<ExecuteArgs> for DeleteVolume {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery
-      .request(api::volume::DeleteVolume {
+      .request(api::docker::DeleteVolume {
         name: self.name.clone(),
       })
       .await
@@ -953,10 +958,10 @@ impl Resolve<ExecuteArgs> for PruneVolumes {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log =
-      match periphery.request(api::volume::PruneVolumes {}).await {
+      match periphery.request(api::docker::PruneVolumes {}).await {
         Ok(log) => log,
         Err(e) => Log::error(
           "prune volumes",
@@ -1005,7 +1010,7 @@ impl Resolve<ExecuteArgs> for PruneDockerBuilders {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log =
       match periphery.request(api::build::PruneBuilders {}).await {
@@ -1057,7 +1062,7 @@ impl Resolve<ExecuteArgs> for PruneBuildx {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log =
       match periphery.request(api::build::PruneBuildx {}).await {
@@ -1109,7 +1114,7 @@ impl Resolve<ExecuteArgs> for PruneSystem {
 
     update_update(update.clone()).await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let log = match periphery.request(api::PruneSystem {}).await {
       Ok(log) => log,

bin/core/src/api/execute/stack.rs

@@ -155,7 +155,8 @@ impl Resolve<ExecuteArgs> for DeployStack {
       compose_config,
       commit_hash,
       commit_message,
-    } = periphery_client(&server)?
+    } = periphery_client(&server)
+      .await?
       .request(ComposeUp {
         stack: stack.clone(),
         services: self.services,
@@ -749,7 +750,8 @@ pub async fn pull_stack_inner(
     Default::default()
   };
 
-  let res = periphery_client(server)?
+  let res = periphery_client(server)
+    .await?
     .request(ComposePull {
       stack,
       services,
@@ -1022,7 +1024,8 @@ impl Resolve<ExecuteArgs> for RunStackService {
       Default::default()
     };
 
-    let log = periphery_client(&server)?
+    let log = periphery_client(&server)
+      .await?
       .request(ComposeRun {
         stack,
         repo,

bin/core/src/api/read/deployment.rs

@@ -145,7 +145,8 @@ impl Resolve<ReadArgs> for GetDeploymentLog {
       return Ok(Log::default());
     }
     let server = resource::get::<Server>(&server_id).await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(api::container::GetContainerLog {
         name,
         tail: cmp::min(tail, MAX_LOG_LENGTH),
@@ -183,7 +184,8 @@ impl Resolve<ReadArgs> for SearchDeploymentLog {
       return Ok(Log::default());
     }
     let server = resource::get::<Server>(&server_id).await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(api::container::GetContainerLogSearch {
         name,
         terms,
@@ -234,7 +236,8 @@ impl Resolve<ReadArgs> for InspectDeploymentContainer {
         .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectContainer { name })
       .await?;
     Ok(res)
@@ -262,7 +265,8 @@ impl Resolve<ReadArgs> for GetDeploymentStats {
       );
     }
     let server = resource::get::<Server>(&server_id).await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(api::container::GetContainerStats { name })
       .await
       .context("failed to get stats from periphery")?;
@@ -321,7 +325,9 @@ impl Resolve<ReadArgs> for GetDeploymentsSummary {
           res.not_deployed += 1;
         }
         DeploymentState::Unknown => {
-          res.unknown += 1;
+          if !deployment.template {
+            res.unknown += 1;
+          }
         }
         _ => {
           res.unhealthy += 1;

bin/core/src/api/read/mod.rs

@@ -27,7 +27,9 @@ use typeshare::typeshare;
 use uuid::Uuid;
 
 use crate::{
-  auth::auth_request, config::core_config, helpers::periphery_client,
+  auth::auth_request,
+  config::{core_config, core_public_key},
+  helpers::periphery_client,
   resource,
 };
 
@@ -39,6 +41,7 @@ mod alerter;
 mod build;
 mod builder;
 mod deployment;
+mod onboarding_key;
 mod permission;
 mod procedure;
 mod provider;
@@ -106,27 +109,29 @@ enum ReadRequest {
   GetServersSummary(GetServersSummary),
   GetServer(GetServer),
   GetServerState(GetServerState),
-  GetPeripheryVersion(GetPeripheryVersion),
+  GetPeripheryInformation(GetPeripheryInformation),
   GetServerActionState(GetServerActionState),
   GetHistoricalServerStats(GetHistoricalServerStats),
   ListServers(ListServers),
   ListFullServers(ListFullServers),
+  ListTerminals(ListTerminals),
+
+  // ==== DOCKER ====
+  GetDockerContainersSummary(GetDockerContainersSummary),
+  ListAllDockerContainers(ListAllDockerContainers),
+  ListDockerContainers(ListDockerContainers),
   InspectDockerContainer(InspectDockerContainer),
   GetResourceMatchingContainer(GetResourceMatchingContainer),
   GetContainerLog(GetContainerLog),
   SearchContainerLog(SearchContainerLog),
+  ListComposeProjects(ListComposeProjects),
+  ListDockerNetworks(ListDockerNetworks),
   InspectDockerNetwork(InspectDockerNetwork),
+  ListDockerImages(ListDockerImages),
   InspectDockerImage(InspectDockerImage),
   ListDockerImageHistory(ListDockerImageHistory),
-  InspectDockerVolume(InspectDockerVolume),
-  GetDockerContainersSummary(GetDockerContainersSummary),
-  ListAllDockerContainers(ListAllDockerContainers),
-  ListDockerContainers(ListDockerContainers),
-  ListDockerNetworks(ListDockerNetworks),
-  ListDockerImages(ListDockerImages),
   ListDockerVolumes(ListDockerVolumes),
-  ListComposeProjects(ListComposeProjects),
-  ListTerminals(ListTerminals),
+  InspectDockerVolume(InspectDockerVolume),
 
   // ==== SERVER STATS ====
   GetSystemInformation(GetSystemInformation),
@@ -224,6 +229,9 @@ enum ReadRequest {
   ListGitProviderAccounts(ListGitProviderAccounts),
   GetDockerRegistryAccount(GetDockerRegistryAccount),
   ListDockerRegistryAccounts(ListDockerRegistryAccounts),
+
+  // ==== ONBOARDING KEY ====
+  ListOnboardingKeys(ListOnboardingKeys),
 }
 
 pub fn router() -> Router {
@@ -298,6 +306,7 @@ fn core_info() -> &'static GetCoreInfoResponse {
         .map(|i| i.namespace.to_string())
         .collect(),
       timezone: config.timezone.clone(),
+      public_key: core_public_key().to_string(),
     }
   })
 }
@@ -343,7 +352,8 @@ impl Resolve<ReadArgs> for ListSecrets {
       };
       if let Some(id) = server_id {
         let server = resource::get::<Server>(&id).await?;
-        let more = periphery_client(&server)?
+        let more = periphery_client(&server)
+          .await?
           .request(periphery_client::api::ListSecrets {})
           .await
           .with_context(|| {
@@ -515,7 +525,8 @@ async fn merge_git_providers_for_server(
   server_id: &str,
 ) -> serror::Result<()> {
   let server = resource::get::<Server>(server_id).await?;
-  let more = periphery_client(&server)?
+  let more = periphery_client(&server)
+    .await?
     .request(periphery_client::api::ListGitProviders {})
     .await
     .with_context(|| {
@@ -553,7 +564,8 @@ async fn merge_docker_registries_for_server(
   server_id: &str,
 ) -> serror::Result<()> {
   let server = resource::get::<Server>(server_id).await?;
-  let more = periphery_client(&server)?
+  let more = periphery_client(&server)
+    .await?
     .request(periphery_client::api::ListDockerRegistries {})
     .await
     .with_context(|| {

bin/core/src/api/read/onboarding_key.rs

@@ -0,0 +1,30 @@
+use anyhow::{Context, anyhow};
+use database::mungos::find::find_collect;
+use komodo_client::api::read::{
+  ListOnboardingKeys, ListOnboardingKeysResponse,
+};
+use reqwest::StatusCode;
+use resolver_api::Resolve;
+use serror::AddStatusCodeError;
+
+use crate::{api::read::ReadArgs, state::db_client};
+
+//
+
+impl Resolve<ReadArgs> for ListOnboardingKeys {
+  async fn resolve(
+    self,
+    ReadArgs { user: admin }: &ReadArgs,
+  ) -> serror::Result<ListOnboardingKeysResponse> {
+    if !admin.admin {
+      return Err(
+        anyhow!("This call is admin only")
+          .status_code(StatusCode::FORBIDDEN),
+      );
+    }
+    find_collect(&db_client().onboarding_keys, None, None)
+      .await
+      .context("Failed to query database for Server onboarding keys")
+      .map_err(Into::into)
+  }
+}

bin/core/src/api/read/repo.rs

@@ -142,7 +142,11 @@ impl Resolve<ReadArgs> for GetReposSummary {
         }
         (RepoState::Ok, _) => res.ok += 1,
         (RepoState::Failed, _) => res.failed += 1,
-        (RepoState::Unknown, _) => res.unknown += 1,
+        (RepoState::Unknown, _) => {
+          if !repo.template {
+            res.unknown += 1
+          }
+        }
         // will never come off the cache in the building state, since that comes from action states
         (RepoState::Cloning, _)
         | (RepoState::Pulling, _)

bin/core/src/api/read/server.rs

@@ -39,18 +39,17 @@ use komodo_client::{
 use periphery_client::api::{
   self as periphery,
   container::InspectContainer,
-  image::{ImageHistory, InspectImage},
-  network::InspectNetwork,
-  volume::InspectVolume,
+  docker::{
+    ImageHistory, InspectImage, InspectNetwork, InspectVolume,
+  },
 };
+use reqwest::StatusCode;
 use resolver_api::Resolve;
+use serror::AddStatusCode;
 use tokio::sync::Mutex;
 
 use crate::{
-  helpers::{
-    periphery_client,
-    query::{get_all_tags, get_system_info},
-  },
+  helpers::{periphery_client, query::get_all_tags},
   permission::get_check_permissions,
   resource,
   stack::compose_container_match_regex,
@@ -80,11 +79,8 @@ impl Resolve<ReadArgs> for GetServersSummary {
       match server.info.state {
         ServerState::Ok => {
           // Check for version mismatch
-          let has_version_mismatch = !server.info.version.is_empty()
-            && server.info.version != "Unknown"
-            && server.info.version != core_version;
-
-          if has_version_mismatch {
+          if matches!(&server.info.version, Some(version) if version != core_version)
+          {
             res.warning += 1;
           } else {
             res.healthy += 1;
@@ -94,7 +90,9 @@ impl Resolve<ReadArgs> for GetServersSummary {
           res.unhealthy += 1;
         }
         ServerState::Disabled => {
-          res.disabled += 1;
+          if !server.template {
+            res.disabled += 1;
+          }
         }
       }
     }
@@ -102,26 +100,6 @@ impl Resolve<ReadArgs> for GetServersSummary {
   }
 }
 
-impl Resolve<ReadArgs> for GetPeripheryVersion {
-  async fn resolve(
-    self,
-    ReadArgs { user }: &ReadArgs,
-  ) -> serror::Result<GetPeripheryVersionResponse> {
-    let server = get_check_permissions::<Server>(
-      &self.server,
-      user,
-      PermissionLevel::Read.into(),
-    )
-    .await?;
-    let version = server_status_cache()
-      .get(&server.id)
-      .await
-      .map(|s| s.version.clone())
-      .unwrap_or(String::from("unknown"));
-    Ok(GetPeripheryVersionResponse { version })
-  }
-}
-
 impl Resolve<ReadArgs> for GetServer {
   async fn resolve(
     self,
@@ -225,6 +203,29 @@ impl Resolve<ReadArgs> for GetServerActionState {
   }
 }
 
+impl Resolve<ReadArgs> for GetPeripheryInformation {
+  async fn resolve(
+    self,
+    ReadArgs { user }: &ReadArgs,
+  ) -> serror::Result<GetPeripheryInformationResponse> {
+    let server = get_check_permissions::<Server>(
+      &self.server,
+      user,
+      PermissionLevel::Read.into(),
+    )
+    .await?;
+    server_status_cache()
+      .get(&server.id)
+      .await
+      .context("Missing server status")?
+      .periphery_info
+      .as_ref()
+      .cloned()
+      .context("Server status missing Periphery Info. The Server may be disconnected.")
+      .status_code(StatusCode::INTERNAL_SERVER_ERROR)
+  }
+}
+
 impl Resolve<ReadArgs> for GetSystemInformation {
   async fn resolve(
     self,
@@ -235,8 +236,17 @@ impl Resolve<ReadArgs> for GetSystemInformation {
       user,
       PermissionLevel::Read.into(),
     )
-    .await?;
-    get_system_info(&server).await.map_err(Into::into)
+    .await
+    .status_code(StatusCode::BAD_REQUEST)?;
+    server_status_cache()
+      .get(&server.id)
+      .await
+      .context("Missing server status")?
+      .system_info
+      .as_ref()
+      .cloned()
+      .context("Server status missing system Info. The Server may be disconnected.")
+      .status_code(StatusCode::INTERNAL_SERVER_ERROR)
   }
 }
 
@@ -251,15 +261,15 @@ impl Resolve<ReadArgs> for GetSystemStats {
       PermissionLevel::Read.into(),
     )
     .await?;
-    let status =
-      server_status_cache().get(&server.id).await.with_context(
-        || format!("did not find status for server at {}", server.id),
-      )?;
-    let stats = status
-      .stats
+    server_status_cache()
+      .get(&server.id)
+      .await
+      .context("Missing server status")?
+      .system_stats
       .as_ref()
-      .context("server stats not available")?;
-    Ok(stats.clone())
+      .cloned()
+      .context("Server status missing system stats. The Server may be disconnected.")
+      .status_code(StatusCode::INTERNAL_SERVER_ERROR)
   }
 }
 
@@ -289,7 +299,8 @@ impl Resolve<ReadArgs> for ListSystemProcesses {
         cached.0.clone()
       }
       _ => {
-        let stats = periphery_client(&server)?
+        let stats = periphery_client(&server)
+          .await?
           .request(periphery::stats::GetSystemProcesses {})
           .await?;
         lock.insert(
@@ -478,7 +489,8 @@ impl Resolve<ReadArgs> for InspectDockerContainer {
         .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectContainer {
         name: self.container,
       })
@@ -506,7 +518,8 @@ impl Resolve<ReadArgs> for GetContainerLog {
       PermissionLevel::Read.logs(),
     )
     .await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(periphery::container::GetContainerLog {
         name: container,
         tail: cmp::min(tail, MAX_LOG_LENGTH),
@@ -537,7 +550,8 @@ impl Resolve<ReadArgs> for SearchContainerLog {
       PermissionLevel::Read.logs(),
     )
     .await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(periphery::container::GetContainerLogSearch {
         name: container,
         terms,
@@ -657,7 +671,8 @@ impl Resolve<ReadArgs> for InspectDockerNetwork {
         .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectNetwork { name: self.network })
       .await?;
     Ok(res)
@@ -706,7 +721,8 @@ impl Resolve<ReadArgs> for InspectDockerImage {
           .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectImage { name: self.image })
       .await?;
     Ok(res)
@@ -736,7 +752,8 @@ impl Resolve<ReadArgs> for ListDockerImageHistory {
         .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(ImageHistory { name: self.image })
       .await?;
     Ok(res)
@@ -785,7 +802,8 @@ impl Resolve<ReadArgs> for InspectDockerVolume {
           .into(),
       );
     }
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectVolume { name: self.volume })
       .await?;
     Ok(res)
@@ -865,7 +883,8 @@ impl Resolve<ReadArgs> for ListTerminals {
     let cache = terminals_cache().get_or_insert(server.id.clone());
     let mut cache = cache.lock().await;
     if self.fresh || komodo_timestamp() > cache.ttl {
-      cache.list = periphery_client(&server)?
+      cache.list = periphery_client(&server)
+        .await?
         .request(periphery_client::api::terminal::ListTerminals {})
         .await
         .context("Failed to get fresh terminal list")?;

bin/core/src/api/read/stack.rs

@@ -89,7 +89,8 @@ impl Resolve<ReadArgs> for GetStackLog {
       true,
     )
     .await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(GetComposeLog {
         project: stack.project_name(false),
         services,
@@ -122,7 +123,8 @@ impl Resolve<ReadArgs> for SearchStackLog {
       true,
     )
     .await?;
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(GetComposeLogSearch {
         project: stack.project_name(false),
         services,
@@ -184,7 +186,8 @@ impl Resolve<ReadArgs> for InspectStackContainer {
         "No service found matching '{service}'. Was the stack last deployed manually?"
       ).into());
     };
-    let res = periphery_client(&server)?
+    let res = periphery_client(&server)
+      .await?
       .request(InspectContainer { name })
       .await?;
     Ok(res)
@@ -363,7 +366,11 @@ impl Resolve<ReadArgs> for GetStacksSummary {
         StackState::Running => res.running += 1,
         StackState::Stopped | StackState::Paused => res.stopped += 1,
         StackState::Down => res.down += 1,
-        StackState::Unknown => res.unknown += 1,
+        StackState::Unknown => {
+          if !stack.template {
+            res.unknown += 1
+          }
+        }
         _ => res.unhealthy += 1,
       }
     }

bin/core/src/api/terminal.rs

@@ -54,34 +54,20 @@ async fn execute_terminal_inner(
 ) -> serror::Result<axum::body::Body> {
   info!("/terminal/execute request | user: {}", user.username);
 
-  let res = async {
-    let server = get_check_permissions::<Server>(
-      &server,
-      &user,
-      PermissionLevel::Read.terminal(),
-    )
-    .await?;
+  let server = get_check_permissions::<Server>(
+    &server,
+    &user,
+    PermissionLevel::Read.terminal(),
+  )
+  .await?;
 
-    let periphery = periphery_client(&server)?;
+  let stream = periphery_client(&server)
+    .await?
+    .execute_terminal(terminal, command)
+    .await
+    .context("Failed to execute command on periphery")?;
 
-    let stream = periphery
-      .execute_terminal(terminal, command)
-      .await
-      .context("Failed to execute command on periphery")?;
-
-    anyhow::Ok(stream)
-  }
-  .await;
-
-  let stream = match res {
-    Ok(stream) => stream,
-    Err(e) => {
-      warn!("/terminal/execute request {req_id} error: {e:#}");
-      return Err(e.into());
-    }
-  };
-
-  Ok(axum::body::Body::from_stream(stream.into_line_stream()))
+  Ok(axum::body::Body::from_stream(stream))
 }
 
 // ======================
@@ -112,43 +98,25 @@ async fn execute_container_exec_inner(
   }: ExecuteContainerExecBody,
   user: User,
 ) -> serror::Result<axum::body::Body> {
-  info!(
-    "/terminal/execute/container request | user: {}",
-    user.username
-  );
+  info!("ExecuteContainerExec request | user: {}", user.username);
 
-  let res = async {
-    let server = get_check_permissions::<Server>(
-      &server,
-      &user,
-      PermissionLevel::Read.terminal(),
-    )
-    .await?;
+  let server = get_check_permissions::<Server>(
+    &server,
+    &user,
+    PermissionLevel::Read.terminal(),
+  )
+  .await?;
 
-    let periphery = periphery_client(&server)?;
+  let periphery = periphery_client(&server).await?;
 
-    let stream = periphery
-      .execute_container_exec(container, shell, command)
-      .await
-      .context(
-        "Failed to execute container exec command on periphery",
-      )?;
+  let stream = periphery
+    .execute_container_exec(container, shell, command)
+    .await
+    .context(
+      "Failed to execute container exec command on periphery",
+    )?;
 
-    anyhow::Ok(stream)
-  }
-  .await;
-
-  let stream = match res {
-    Ok(stream) => stream,
-    Err(e) => {
-      warn!(
-        "/terminal/execute/container request {req_id} error: {e:#}"
-      );
-      return Err(e.into());
-    }
-  };
-
-  Ok(axum::body::Body::from_stream(stream.into_line_stream()))
+  Ok(axum::body::Body::from_stream(stream))
 }
 
 // =======================
@@ -178,45 +146,27 @@ async fn execute_deployment_exec_inner(
   }: ExecuteDeploymentExecBody,
   user: User,
 ) -> serror::Result<axum::body::Body> {
-  info!(
-    "/terminal/execute/deployment request | user: {}",
-    user.username
-  );
+  info!("ExecuteDeploymentExec request | user: {}", user.username);
 
-  let res = async {
-    let deployment = get_check_permissions::<Deployment>(
-      &deployment,
-      &user,
-      PermissionLevel::Read.terminal(),
-    )
-    .await?;
+  let deployment = get_check_permissions::<Deployment>(
+    &deployment,
+    &user,
+    PermissionLevel::Read.terminal(),
+  )
+  .await?;
 
-    let server = get::<Server>(&deployment.config.server_id).await?;
+  let server = get::<Server>(&deployment.config.server_id).await?;
 
-    let periphery = periphery_client(&server)?;
+  let periphery = periphery_client(&server).await?;
 
-    let stream = periphery
-      .execute_container_exec(deployment.name, shell, command)
-      .await
-      .context(
-        "Failed to execute container exec command on periphery",
-      )?;
+  let stream = periphery
+    .execute_container_exec(deployment.name, shell, command)
+    .await
+    .context(
+      "Failed to execute container exec command on periphery",
+    )?;
 
-    anyhow::Ok(stream)
-  }
-  .await;
-
-  let stream = match res {
-    Ok(stream) => stream,
-    Err(e) => {
-      warn!(
-        "/terminal/execute/deployment request {req_id} error: {e:#}"
-      );
-      return Err(e.into());
-    }
-  };
-
-  Ok(axum::body::Body::from_stream(stream.into_line_stream()))
+  Ok(axum::body::Body::from_stream(stream))
 }
 
 // ==================
@@ -247,53 +197,40 @@ async fn execute_stack_exec_inner(
   }: ExecuteStackExecBody,
   user: User,
 ) -> serror::Result<axum::body::Body> {
-  info!("/terminal/execute/stack request | user: {}", user.username);
+  info!("ExecuteStackExec request | user: {}", user.username);
 
-  let res = async {
-    let stack = get_check_permissions::<Stack>(
-      &stack,
-      &user,
-      PermissionLevel::Read.terminal(),
-    )
-    .await?;
+  let stack = get_check_permissions::<Stack>(
+    &stack,
+    &user,
+    PermissionLevel::Read.terminal(),
+  )
+  .await?;
 
-    let server = get::<Server>(&stack.config.server_id).await?;
+  let server = get::<Server>(&stack.config.server_id).await?;
 
-    let container = stack_status_cache()
-      .get(&stack.id)
-      .await
-      .context("could not get stack status")?
-      .curr
-      .services
-      .iter()
-      .find(|s| s.service == service)
-      .context("could not find service")?
-      .container
-      .as_ref()
-      .context("could not find service container")?
-      .name
-      .clone();
+  let container = stack_status_cache()
+    .get(&stack.id)
+    .await
+    .context("could not get stack status")?
+    .curr
+    .services
+    .iter()
+    .find(|s| s.service == service)
+    .context("could not find service")?
+    .container
+    .as_ref()
+    .context("could not find service container")?
+    .name
+    .clone();
 
-    let periphery = periphery_client(&server)?;
+  let periphery = periphery_client(&server).await?;
 
-    let stream = periphery
-      .execute_container_exec(container, shell, command)
-      .await
-      .context(
-        "Failed to execute container exec command on periphery",
-      )?;
+  let stream = periphery
+    .execute_container_exec(container, shell, command)
+    .await
+    .context(
+      "Failed to execute container exec command on periphery",
+    )?;
 
-    anyhow::Ok(stream)
-  }
-  .await;
-
-  let stream = match res {
-    Ok(stream) => stream,
-    Err(e) => {
-      warn!("/terminal/execute/stack request {req_id} error: {e:#}");
-      return Err(e.into());
-    }
-  };
-
-  Ok(axum::body::Body::from_stream(stream.into_line_stream()))
+  Ok(axum::body::Body::from_stream(stream))
 }

bin/core/src/api/write/action.rs

@@ -16,7 +16,8 @@ impl Resolve<WriteArgs> for CreateAction {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Action> {
-    resource::create::<Action>(&self.name, self.config, user).await
+    resource::create::<Action>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -32,7 +33,8 @@ impl Resolve<WriteArgs> for CopyAction {
       PermissionLevel::Write.into(),
     )
     .await?;
-    resource::create::<Action>(&self.name, config.into(), user).await
+    resource::create::<Action>(&self.name, config.into(), None, user)
+      .await
   }
 }
 

bin/core/src/api/write/alerter.rs

@@ -16,7 +16,8 @@ impl Resolve<WriteArgs> for CreateAlerter {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Alerter> {
-    resource::create::<Alerter>(&self.name, self.config, user).await
+    resource::create::<Alerter>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -32,7 +33,8 @@ impl Resolve<WriteArgs> for CopyAlerter {
       PermissionLevel::Write.into(),
     )
     .await?;
-    resource::create::<Alerter>(&self.name, config.into(), user).await
+    resource::create::<Alerter>(&self.name, config.into(), None, user)
+      .await
   }
 }
 

bin/core/src/api/write/build.rs

@@ -1,8 +1,10 @@
 use std::{path::PathBuf, str::FromStr, time::Duration};
 
 use anyhow::{Context, anyhow};
-use database::mongo_indexed::doc;
 use database::mungos::mongodb::bson::to_document;
+use database::{
+  mongo_indexed::doc, mungos::mongodb::bson::oid::ObjectId,
+};
 use formatting::format_serror;
 use komodo_client::{
   api::write::*,
@@ -21,22 +23,21 @@ use komodo_client::{
 use octorust::types::{
   ReposCreateWebhookRequest, ReposCreateWebhookRequestConfig,
 };
-use periphery_client::{
-  PeripheryClient,
-  api::build::{
-    GetDockerfileContentsOnHost, WriteDockerfileContentsToHost,
-  },
+use periphery_client::api::build::{
+  GetDockerfileContentsOnHost, WriteDockerfileContentsToHost,
 };
 use resolver_api::Resolve;
 use tokio::fs;
 
 use crate::{
   config::core_config,
+  connection::PeripheryConnectionArgs,
   helpers::{
     git_token, periphery_client,
     query::get_server_with_state,
     update::{add_update, make_update},
   },
+  periphery::PeripheryClient,
   permission::get_check_permissions,
   resource,
   state::{db_client, github_client},
@@ -50,7 +51,8 @@ impl Resolve<WriteArgs> for CreateBuild {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Build> {
-    resource::create::<Build>(&self.name, self.config, user).await
+    resource::create::<Build>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -68,7 +70,8 @@ impl Resolve<WriteArgs> for CopyBuild {
     .await?;
     // reset version to 0.0.0
     config.version = Default::default();
-    resource::create::<Build>(&self.name, config.into(), user).await
+    resource::create::<Build>(&self.name, config.into(), None, user)
+      .await
   }
 }
 
@@ -428,13 +431,27 @@ async fn get_on_host_periphery(
       Err(anyhow!("Files on host doesn't work with AWS builder"))
     }
     BuilderConfig::Url(config) => {
+      // TODO: Ensure connection is actually established.
+      // Builder id no good because it may be active for multiple connections.
       let periphery = PeripheryClient::new(
-        config.address,
-        config.passkey,
-        Duration::from_secs(3),
-      );
-      periphery.health_check().await?;
-      Ok(periphery)
+        PeripheryConnectionArgs::from_url_builder(
+          &ObjectId::new().to_hex(),
+          &config,
+        ),
+        config.insecure_tls,
+        &config.passkey,
+      )
+      .await?;
+      // Poll for connection to be estalished
+      let mut err = None;
+      for _ in 0..10 {
+        tokio::time::sleep(Duration::from_secs(1)).await;
+        match periphery.health_check().await {
+          Ok(_) => return Ok(periphery),
+          Err(e) => err = Some(e),
+        };
+      }
+      Err(err.context("Missing error")?)
     }
     BuilderConfig::Server(config) => {
       if config.server_id.is_empty() {
@@ -449,7 +466,7 @@ async fn get_on_host_periphery(
           "Builder server is disabled or not reachable"
         ));
       };
-      periphery_client(&server)
+      periphery_client(&server).await
     }
   }
 }

bin/core/src/api/write/builder.rs

@@ -16,7 +16,8 @@ impl Resolve<WriteArgs> for CreateBuilder {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Builder> {
-    resource::create::<Builder>(&self.name, self.config, user).await
+    resource::create::<Builder>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -32,7 +33,8 @@ impl Resolve<WriteArgs> for CopyBuilder {
       PermissionLevel::Write.into(),
     )
     .await?;
-    resource::create::<Builder>(&self.name, config.into(), user).await
+    resource::create::<Builder>(&self.name, config.into(), None, user)
+      .await
   }
 }
 

bin/core/src/api/write/deployment.rs

@@ -38,8 +38,13 @@ impl Resolve<WriteArgs> for CreateDeployment {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Deployment> {
-    resource::create::<Deployment>(&self.name, self.config, user)
-      .await
+    resource::create::<Deployment>(
+      &self.name,
+      self.config,
+      None,
+      user,
+    )
+    .await
   }
 }
 
@@ -56,8 +61,13 @@ impl Resolve<WriteArgs> for CopyDeployment {
         PermissionLevel::Read.into(),
       )
       .await?;
-    resource::create::<Deployment>(&self.name, config.into(), user)
-      .await
+    resource::create::<Deployment>(
+      &self.name,
+      config.into(),
+      None,
+      user,
+    )
+    .await
   }
 }
 
@@ -85,7 +95,8 @@ impl Resolve<WriteArgs> for CreateDeploymentFromContainer {
         .into(),
       );
     }
-    let container = periphery_client(&server)?
+    let container = periphery_client(&server)
+      .await?
       .request(InspectContainer {
         name: self.name.clone(),
       })
@@ -149,7 +160,8 @@ impl Resolve<WriteArgs> for CreateDeploymentFromContainer {
       });
     }
 
-    resource::create::<Deployment>(&self.name, config, user).await
+    resource::create::<Deployment>(&self.name, config, None, user)
+      .await
   }
 }
 
@@ -231,7 +243,8 @@ impl Resolve<WriteArgs> for RenameDeployment {
     if container_state != DeploymentState::NotDeployed {
       let server =
         resource::get::<Server>(&deployment.config.server_id).await?;
-      let log = periphery_client(&server)?
+      let log = periphery_client(&server)
+        .await?
         .request(api::container::RenameContainer {
           curr_name: deployment.name.clone(),
           new_name: name.clone(),

bin/core/src/api/write/mod.rs

@@ -23,6 +23,7 @@ mod alerter;
 mod build;
 mod builder;
 mod deployment;
+mod onboarding_key;
 mod permissions;
 mod procedure;
 mod provider;
@@ -91,6 +92,8 @@ pub enum WriteRequest {
   CreateTerminal(CreateTerminal),
   DeleteTerminal(DeleteTerminal),
   DeleteAllTerminals(DeleteAllTerminals),
+  UpdateServerPublicKey(UpdateServerPublicKey),
+  RotateServerKeys(RotateServerKeys),
 
   // ==== STACK ====
   CreateStack(CreateStack),
@@ -185,13 +188,18 @@ pub enum WriteRequest {
   UpdateVariableIsSecret(UpdateVariableIsSecret),
   DeleteVariable(DeleteVariable),
 
-  // ==== PROVIDERS ====
+  // ==== PROVIDER ====
   CreateGitProviderAccount(CreateGitProviderAccount),
   UpdateGitProviderAccount(UpdateGitProviderAccount),
   DeleteGitProviderAccount(DeleteGitProviderAccount),
   CreateDockerRegistryAccount(CreateDockerRegistryAccount),
   UpdateDockerRegistryAccount(UpdateDockerRegistryAccount),
   DeleteDockerRegistryAccount(DeleteDockerRegistryAccount),
+
+  // ==== ONBOARDING KEY ====
+  CreateOnboardingKey(CreateOnboardingKey),
+  UpdateOnboardingKey(UpdateOnboardingKey),
+  DeleteOnboardingKey(DeleteOnboardingKey),
 }
 
 pub fn router() -> Router {

bin/core/src/api/write/onboarding_key.rs

@@ -0,0 +1,168 @@
+use anyhow::{Context, anyhow};
+use database::mungos::mongodb::bson::{Document, doc};
+use komodo_client::{
+  api::write::{
+    CreateOnboardingKey, CreateOnboardingKeyResponse,
+    DeleteOnboardingKey, DeleteOnboardingKeyResponse,
+    UpdateOnboardingKey, UpdateOnboardingKeyResponse,
+  },
+  entities::{komodo_timestamp, onboarding_key::OnboardingKey},
+};
+use noise::key::EncodedKeyPair;
+use reqwest::StatusCode;
+use resolver_api::Resolve;
+use serror::{AddStatusCode, AddStatusCodeError};
+
+use crate::{api::write::WriteArgs, state::db_client};
+
+//
+
+impl Resolve<WriteArgs> for CreateOnboardingKey {
+  #[instrument(name = "CreateServerOnboardingKey", skip(self, admin))]
+  async fn resolve(
+    self,
+    WriteArgs { user: admin }: &WriteArgs,
+  ) -> serror::Result<CreateOnboardingKeyResponse> {
+    if !admin.admin {
+      return Err(
+        anyhow!("This call is admin only")
+          .status_code(StatusCode::FORBIDDEN),
+      );
+    }
+    let keys = if let Some(private_key) = self.private_key {
+      EncodedKeyPair::from_private_key(&private_key)?
+    } else {
+      EncodedKeyPair::generate()?
+    };
+    let onboarding_key = OnboardingKey {
+      public_key: keys.public.into_inner(),
+      name: self.name,
+      enabled: true,
+      onboarded: Default::default(),
+      created_at: komodo_timestamp(),
+      expires: self.expires,
+      tags: self.tags,
+      copy_server: self.copy_server,
+      create_builder: self.create_builder,
+    };
+    let db = db_client();
+    // Create the key
+    db.onboarding_keys
+      .insert_one(&onboarding_key)
+      .await
+      .context(
+        "Failed to create Server onboarding key on database",
+      )?;
+    let created = db
+      .onboarding_keys
+      .find_one(doc! { "public_key": &onboarding_key.public_key })
+      .await
+      .context("Failed to query database for Server onboarding keys")?
+      .context(
+        "No Server onboarding key found on database after create",
+      )?;
+    Ok(CreateOnboardingKeyResponse {
+      private_key: keys.private.into_inner(),
+      created,
+    })
+  }
+}
+
+//
+
+impl Resolve<WriteArgs> for UpdateOnboardingKey {
+  async fn resolve(
+    self,
+    WriteArgs { user: admin }: &WriteArgs,
+  ) -> serror::Result<UpdateOnboardingKeyResponse> {
+    if !admin.admin {
+      return Err(
+        anyhow!("This call is admin only")
+          .status_code(StatusCode::FORBIDDEN),
+      );
+    }
+
+    let query = doc! { "public_key": &self.public_key };
+
+    // No changes
+    if self.is_none() {
+      return db_client()
+        .onboarding_keys
+        .find_one(query)
+        .await
+        .context("Failed to query database for onboarding key")?
+        .context("No matching onboarding key found")
+        .status_code(StatusCode::NOT_FOUND);
+    }
+
+    let mut update = Document::new();
+
+    if let Some(enabled) = self.enabled {
+      update.insert("enabled", enabled);
+    }
+
+    if let Some(name) = self.name {
+      update.insert("name", name);
+    }
+
+    if let Some(expires) = self.expires {
+      update.insert("expires", expires);
+    }
+
+    if let Some(tags) = self.tags {
+      update.insert("tags", tags);
+    }
+
+    if let Some(copy_server) = self.copy_server {
+      update.insert("copy_server", copy_server);
+    }
+
+    if let Some(create_builder) = self.create_builder {
+      update.insert("create_builder", create_builder);
+    }
+
+    db_client()
+      .onboarding_keys
+      .update_one(query.clone(), doc! { "$set": update })
+      .await
+      .context("Failed to update onboarding key on database")?;
+
+    db_client()
+      .onboarding_keys
+      .find_one(query)
+      .await
+      .context("Failed to query database for onboarding key")?
+      .context("No matching onboarding key found")
+      .status_code(StatusCode::NOT_FOUND)
+  }
+}
+
+//
+
+impl Resolve<WriteArgs> for DeleteOnboardingKey {
+  #[instrument(name = "DeleteServerOnboardingKey", skip(admin))]
+  async fn resolve(
+    self,
+    WriteArgs { user: admin }: &WriteArgs,
+  ) -> serror::Result<DeleteOnboardingKeyResponse> {
+    if !admin.admin {
+      return Err(
+        anyhow!("This call is admin only")
+          .status_code(StatusCode::FORBIDDEN),
+      );
+    }
+    let db = db_client();
+    let query = doc! { "public_key": &self.public_key };
+    let creation_key = db
+      .onboarding_keys
+      .find_one(query.clone())
+      .await
+      .context("Failed to query database for Server onboarding keys")?
+      .context("Server onboarding key matching provided public key not found")
+      .status_code(StatusCode::NOT_FOUND)?;
+    db.onboarding_keys.delete_one(query).await.context(
+      "Failed to delete Server onboarding key from database",
+    )?;
+    Ok(creation_key)
+  }
+}

bin/core/src/api/write/procedure.rs

@@ -16,7 +16,8 @@ impl Resolve<WriteArgs> for CreateProcedure {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<CreateProcedureResponse> {
-    resource::create::<Procedure>(&self.name, self.config, user).await
+    resource::create::<Procedure>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -33,8 +34,13 @@ impl Resolve<WriteArgs> for CopyProcedure {
         PermissionLevel::Write.into(),
       )
       .await?;
-    resource::create::<Procedure>(&self.name, config.into(), user)
-      .await
+    resource::create::<Procedure>(
+      &self.name,
+      config.into(),
+      None,
+      user,
+    )
+    .await
   }
 }
 

bin/core/src/api/write/repo.rs

@@ -42,7 +42,8 @@ impl Resolve<WriteArgs> for CreateRepo {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Repo> {
-    resource::create::<Repo>(&self.name, self.config, user).await
+    resource::create::<Repo>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -58,7 +59,8 @@ impl Resolve<WriteArgs> for CopyRepo {
       PermissionLevel::Read.into(),
     )
     .await?;
-    resource::create::<Repo>(&self.name, config.into(), user).await
+    resource::create::<Repo>(&self.name, config.into(), None, user)
+      .await
   }
 }
 
@@ -127,7 +129,8 @@ impl Resolve<WriteArgs> for RenameRepo {
     let server =
       resource::get::<Server>(&repo.config.server_id).await?;
 
-    let log = match periphery_client(&server)?
+    let log = match periphery_client(&server)
+      .await?
       .request(api::git::RenameRepo {
         curr_name: to_path_compatible_name(&repo.name),
         new_name: name.clone(),

bin/core/src/api/write/server.rs

@@ -1,11 +1,11 @@
 use anyhow::Context;
-use formatting::format_serror;
+use formatting::{bold, format_serror};
 use komodo_client::{
   api::write::*,
   entities::{
     NoData, Operation,
     permission::PermissionLevel,
-    server::Server,
+    server::{Server, ServerInfo},
     to_docker_compatible_name,
     update::{Update, UpdateStatus},
   },
@@ -19,7 +19,7 @@ use crate::{
     update::{add_update, make_update, update_update},
   },
   permission::get_check_permissions,
-  resource,
+  resource::{self, update_server_public_key},
 };
 
 use super::WriteArgs;
@@ -30,7 +30,16 @@ impl Resolve<WriteArgs> for CreateServer {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Server> {
-    resource::create::<Server>(&self.name, self.config, user).await
+    resource::create::<Server>(
+      &self.name,
+      self.config,
+      self.public_key.map(|public_key| ServerInfo {
+        public_key,
+        ..Default::default()
+      }),
+      user,
+    )
+    .await
   }
 }
 
@@ -47,7 +56,16 @@ impl Resolve<WriteArgs> for CopyServer {
     )
     .await?;
 
-    resource::create::<Server>(&self.name, config.into(), user).await
+    resource::create::<Server>(
+      &self.name,
+      config.into(),
+      self.public_key.map(|public_key| ServerInfo {
+        public_key,
+        ..Default::default()
+      }),
+      user,
+    )
+    .await
   }
 }
 
@@ -91,7 +109,7 @@ impl Resolve<WriteArgs> for CreateNetwork {
     )
     .await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     let mut update =
       make_update(&server, Operation::CreateNetwork, user);
@@ -99,7 +117,7 @@ impl Resolve<WriteArgs> for CreateNetwork {
     update.id = add_update(update.clone()).await?;
 
     match periphery
-      .request(api::network::CreateNetwork {
+      .request(api::docker::CreateNetwork {
         name: to_docker_compatible_name(&self.name),
         driver: None,
       })
@@ -108,7 +126,7 @@ impl Resolve<WriteArgs> for CreateNetwork {
       Ok(log) => update.logs.push(log),
       Err(e) => update.push_error_log(
         "create network",
-        format_serror(&e.context("failed to create network").into()),
+        format_serror(&e.context("Failed to create network").into()),
       ),
     };
 
@@ -132,7 +150,7 @@ impl Resolve<WriteArgs> for CreateTerminal {
     )
     .await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     periphery
       .request(api::terminal::CreateTerminal {
@@ -141,7 +159,7 @@ impl Resolve<WriteArgs> for CreateTerminal {
         recreate: self.recreate,
       })
       .await
-      .context("Failed to create terminal on periphery")?;
+      .context("Failed to create terminal on Periphery")?;
 
     Ok(NoData {})
   }
@@ -160,14 +178,14 @@ impl Resolve<WriteArgs> for DeleteTerminal {
     )
     .await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     periphery
       .request(api::terminal::DeleteTerminal {
         terminal: self.terminal,
       })
       .await
-      .context("Failed to delete terminal on periphery")?;
+      .context("Failed to delete terminal on Periphery")?;
 
     Ok(NoData {})
   }
@@ -186,13 +204,76 @@ impl Resolve<WriteArgs> for DeleteAllTerminals {
     )
     .await?;
 
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     periphery
       .request(api::terminal::DeleteAllTerminals {})
       .await
-      .context("Failed to delete all terminals on periphery")?;
+      .context("Failed to delete all terminals on Periphery")?;
 
     Ok(NoData {})
   }
 }
+
+//
+
+impl Resolve<WriteArgs> for UpdateServerPublicKey {
+  #[instrument(name = "UpdateServerPublicKey", skip(args))]
+  async fn resolve(
+    self,
+    args: &WriteArgs,
+  ) -> Result<Self::Response, Self::Error> {
+    let server = get_check_permissions::<Server>(
+      &self.server,
+      &args.user,
+      PermissionLevel::Write.into(),
+    )
+    .await?;
+
+    update_server_public_key(&server.id, &self.public_key).await?;
+
+    let mut update =
+      make_update(&server, Operation::UpdateServerKey, &args.user);
+
+    update.push_simple_log(
+      "Update Server Public Key",
+      format!("Public key updated to {}", bold(&self.public_key)),
+    );
+    update.finalize();
+    update.id = add_update(update.clone()).await?;
+
+    Ok(update)
+  }
+}
+
+//
+
+impl Resolve<WriteArgs> for RotateServerKeys {
+  #[instrument(name = "RotateServerPrivateKey", skip(args))]
+  async fn resolve(
+    self,
+    args: &WriteArgs,
+  ) -> Result<Self::Response, Self::Error> {
+    let server = get_check_permissions::<Server>(
+      &self.server,
+      &args.user,
+      PermissionLevel::Write.into(),
+    )
+    .await?;
+
+    let periphery = periphery_client(&server).await?;
+
+    let public_key = periphery
+      .request(api::keys::RotatePrivateKey {})
+      .await
+      .context("Failed to rotate Periphery private key")?
+      .public_key;
+
+    UpdateServerPublicKey {
+      server: server.id,
+      public_key,
+    }
+    .resolve(args)
+    .await
+  }
+}

bin/core/src/api/write/stack.rs

@@ -51,7 +51,8 @@ impl Resolve<WriteArgs> for CreateStack {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<Stack> {
-    resource::create::<Stack>(&self.name, self.config, user).await
+    resource::create::<Stack>(&self.name, self.config, None, user)
+      .await
   }
 }
 
@@ -68,7 +69,8 @@ impl Resolve<WriteArgs> for CopyStack {
     )
     .await?;
 
-    resource::create::<Stack>(&self.name, config.into(), user).await
+    resource::create::<Stack>(&self.name, config.into(), None, user)
+      .await
   }
 }
 
@@ -170,7 +172,8 @@ async fn write_stack_file_contents_on_host(
       .into(),
     );
   }
-  match periphery_client(&server)?
+  match periphery_client(&server)
+    .await?
     .request(WriteComposeContentsToHost {
       name: stack.name,
       run_directory: stack.config.run_directory,
@@ -424,7 +427,8 @@ impl Resolve<WriteArgs> for RefreshStackCache {
         (vec![], None, None, None, None)
       } else if let Some(server) = server {
         let GetComposeContentsOnHostResponse { contents, errors } =
-          match periphery_client(&server)?
+          match periphery_client(&server)
+            .await?
             .request(GetComposeContentsOnHost {
               file_paths: stack.all_file_dependencies(),
               name: stack.name.clone(),

bin/core/src/api/write/sync.rs

@@ -68,8 +68,13 @@ impl Resolve<WriteArgs> for CreateResourceSync {
     self,
     WriteArgs { user }: &WriteArgs,
   ) -> serror::Result<ResourceSync> {
-    resource::create::<ResourceSync>(&self.name, self.config, user)
-      .await
+    resource::create::<ResourceSync>(
+      &self.name,
+      self.config,
+      None,
+      user,
+    )
+    .await
   }
 }
 
@@ -86,8 +91,13 @@ impl Resolve<WriteArgs> for CopyResourceSync {
         PermissionLevel::Write.into(),
       )
       .await?;
-    resource::create::<ResourceSync>(&self.name, config.into(), user)
-      .await
+    resource::create::<ResourceSync>(
+      &self.name,
+      config.into(),
+      None,
+      user,
+    )
+    .await
   }
 }
 

bin/core/src/api/write/user_group.rs

@@ -26,7 +26,7 @@ impl Resolve<WriteArgs> for CreateUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -64,7 +64,7 @@ impl Resolve<WriteArgs> for RenameUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -93,7 +93,7 @@ impl Resolve<WriteArgs> for DeleteUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -129,7 +129,7 @@ impl Resolve<WriteArgs> for AddUserToUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -176,7 +176,7 @@ impl Resolve<WriteArgs> for RemoveUserFromUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -223,7 +223,7 @@ impl Resolve<WriteArgs> for SetUsersInUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }
@@ -273,7 +273,7 @@ impl Resolve<WriteArgs> for SetEveryoneUserGroup {
   ) -> serror::Result<UserGroup> {
     if !admin.admin {
       return Err(
-        anyhow!("This call is admin-only")
+        anyhow!("This call is admin only")
           .status_code(StatusCode::FORBIDDEN),
       );
     }

bin/core/src/cloud/aws/ec2.rs

@@ -84,6 +84,8 @@ pub async fn launch_ec2_instance(
     assign_public_ip,
     use_public_ip,
     user_data,
+    periphery_public_key: _,
+    insecure_tls: _,
     port: _,
     use_https: _,
     git_providers: _,

bin/core/src/cloud/mod.rs

@@ -4,6 +4,8 @@ pub mod aws;
 pub enum BuildCleanupData {
   /// Nothing to clean up
   Server,
+  /// Cleanup Periphery connection
+  Url,
   /// Clean up AWS instance
   Aws { instance_id: String, region: String },
 }

bin/core/src/config.rs

@@ -16,6 +16,94 @@ use komodo_client::entities::{
   },
   logger::LogConfig,
 };
+use noise::key::{SpkiPublicKey, load_maybe_generate_private_key};
+
+/// Should call in startup to ensure Core errors without valid private key.
+pub fn core_private_key() -> &'static String {
+  static CORE_PRIVATE_KEY: OnceLock<String> = OnceLock::new();
+  CORE_PRIVATE_KEY.get_or_init(|| {
+    let config = core_config();
+    if let Some(path) = config.private_key.strip_prefix("file:") {
+      load_maybe_generate_private_key(path).unwrap()
+    } else {
+      config.private_key.clone()
+    }
+  })
+}
+
+/// Should call in startup to ensure Core errors without valid private key.
+pub fn core_public_key() -> &'static String {
+  static CORE_PUBLIC_KEY: OnceLock<String> = OnceLock::new();
+  CORE_PUBLIC_KEY.get_or_init(|| {
+    SpkiPublicKey::from_private_key(core_private_key())
+      .context("Got invalid private key")
+      .unwrap()
+      .into_inner()
+  })
+}
+
+pub fn core_connection_query() -> &'static String {
+  static CORE_HOSTNAME: OnceLock<String> = OnceLock::new();
+  CORE_HOSTNAME.get_or_init(|| {
+    let host = url::Url::parse(&core_config().host)
+      .context("Failed to parse config field 'host' as URL")
+      .unwrap()
+      .host()
+      .context(
+        "Failed to parse config field 'host' | missing host part",
+      )
+      .unwrap()
+      .to_string();
+    format!("core={}", urlencoding::encode(&host))
+  })
+}
+
+pub fn periphery_public_keys() -> Option<&'static [SpkiPublicKey]> {
+  static PERIPHERY_PUBLIC_KEYS: OnceLock<Option<Vec<SpkiPublicKey>>> =
+    OnceLock::new();
+  PERIPHERY_PUBLIC_KEYS
+    .get_or_init(|| {
+      core_config().periphery_public_keys.as_ref().map(
+        |public_keys| {
+          public_keys
+            .iter()
+            .flat_map(|public_key| {
+              let (path, maybe_pem) = if let Some(path) =
+                public_key.strip_prefix("file:")
+              {
+                match std::fs::read_to_string(path).with_context(
+                  || format!("Failed to read periphery public key at {path:?}"),
+                ) {
+                  Ok(public_key) => (Some(path), public_key),
+                  Err(e) => {
+                    warn!("{e:#}");
+                    return None;
+                  }
+                }
+              } else {
+                (None, public_key.clone())
+              };
+              match SpkiPublicKey::from_maybe_pem(&maybe_pem) {
+                Ok(public_key) => Some(public_key),
+                Err(e) => {
+                  warn!(
+                    "Failed to read periphery public key{} | {e:#}",
+                    if let Some(path) = path {
+                      format!("at {path:?}")
+                    } else {
+                      String::new()
+                    }
+                  );
+                  None
+                }
+              }
+            })
+            .collect()
+        },
+      )
+    })
+    .as_deref()
+}
 
 pub fn core_config() -> &'static CoreConfig {
   static CORE_CONFIG: OnceLock<CoreConfig> = OnceLock::new();
@@ -88,9 +176,11 @@ pub fn core_config() -> &'static CoreConfig {
     // recreating CoreConfig here makes sure apply all env overrides applied.
     CoreConfig {
       // Secret things overridden with file
-      jwt_secret: maybe_read_item_from_file(env.komodo_jwt_secret_file, env.komodo_jwt_secret).unwrap_or(config.jwt_secret),
+      private_key: maybe_read_item_from_file(env.komodo_private_key_file, env.komodo_private_key)
+        .unwrap_or(config.private_key),
       passkey: maybe_read_item_from_file(env.komodo_passkey_file, env.komodo_passkey)
-        .unwrap_or(config.passkey),
+        .or(config.passkey),
+      jwt_secret: maybe_read_item_from_file(env.komodo_jwt_secret_file, env.komodo_jwt_secret).unwrap_or(config.jwt_secret),
       webhook_secret: maybe_read_item_from_file(env.komodo_webhook_secret_file, env.komodo_webhook_secret)
         .unwrap_or(config.webhook_secret),
       database: DatabaseConfig {
@@ -177,8 +267,9 @@ pub fn core_config() -> &'static CoreConfig {
       port: env.komodo_port.unwrap_or(config.port),
       bind_ip: env.komodo_bind_ip.unwrap_or(config.bind_ip),
       timezone: env.komodo_timezone.unwrap_or(config.timezone),
-      first_server: env.komodo_first_server.or(config.first_server),
-      first_server_name: env.komodo_first_server_name.unwrap_or(config.first_server_name),
+      periphery_public_keys: env.komodo_periphery_public_keys.or(config.periphery_public_keys),
+      first_server_address: env.komodo_first_server_address.or(config.first_server_address),
+      first_server_name: env.komodo_first_server_name.or(config.first_server_name),
       frontend_path: env.komodo_frontend_path.unwrap_or(config.frontend_path),
       jwt_ttl: env
         .komodo_jwt_ttl

bin/core/src/connection/client.rs

@@ -0,0 +1,186 @@
+use std::{sync::Arc, time::Duration};
+
+use anyhow::{Context, anyhow};
+use periphery_client::CONNECTION_RETRY_SECONDS;
+use serror::{deserialize_error_bytes, serialize_error_bytes};
+use transport::{
+  MessageState,
+  auth::{
+    AddressConnectionIdentifiers, ClientLoginFlow,
+    ConnectionIdentifiers,
+  },
+  fix_ws_address,
+  websocket::{Websocket, tungstenite::TungsteniteWebsocket},
+};
+
+use crate::{
+  config::{core_config, core_connection_query},
+  periphery::ConnectionChannels,
+  state::periphery_connections,
+};
+
+use super::{PeripheryConnection, PeripheryConnectionArgs};
+
+impl PeripheryConnectionArgs<'_> {
+  pub async fn spawn_client_connection(
+    self,
+    id: String,
+    insecure: bool,
+    passkey: String,
+  ) -> anyhow::Result<Arc<ConnectionChannels>> {
+    let Some(address) = self.address else {
+      return Err(anyhow!(
+        "Cannot spawn client connection with empty address"
+      ));
+    };
+
+    let address = fix_ws_address(address);
+    let identifiers =
+      AddressConnectionIdentifiers::extract(&address)?;
+    let endpoint = format!("{address}/?{}", core_connection_query());
+
+    let (connection, mut receiver) =
+      periphery_connections().insert(id.clone(), self).await;
+
+    let channels = connection.channels.clone();
+
+    tokio::spawn(async move {
+      loop {
+        let ws = tokio::select! {
+          ws = TungsteniteWebsocket::connect_maybe_tls_insecure(
+            &endpoint,
+            insecure && endpoint.starts_with("wss"),
+          ) => ws,
+          _ = connection.cancel.cancelled() => {
+            break
+          }
+        };
+
+        let (mut socket, accept) = match ws {
+          Ok(res) => res,
+          Err(e) => {
+            connection.set_error(e.error).await;
+            tokio::time::sleep(Duration::from_secs(
+              CONNECTION_RETRY_SECONDS,
+            ))
+            .await;
+            continue;
+          }
+        };
+
+        let identifiers = identifiers.build(
+          accept.as_bytes(),
+          core_connection_query().as_bytes(),
+        );
+
+        if let Err(e) = connection
+          .client_login(&mut socket, identifiers, &passkey)
+          .await
+        {
+          connection.set_error(e).await;
+          tokio::time::sleep(Duration::from_secs(
+            CONNECTION_RETRY_SECONDS,
+          ))
+          .await;
+          continue;
+        };
+
+        connection.handle_socket(socket, &mut receiver).await
+      }
+    });
+
+    Ok(channels)
+  }
+}
+
+impl PeripheryConnection {
+  /// Custom Core -> Periphery side only login wrapper
+  /// to implement passkey support for backward compatibility
+  async fn client_login(
+    &self,
+    socket: &mut TungsteniteWebsocket,
+    identifiers: ConnectionIdentifiers<'_>,
+    // for legacy auth
+    passkey: &str,
+  ) -> anyhow::Result<()> {
+    // Get the required auth type
+    let bytes = socket
+      .recv_bytes()
+      .with_timeout(Duration::from_secs(2))
+      .await?
+      .context("Failed to receive login type indicator")?;
+
+    match bytes.iter().as_slice() {
+      // Noise auth
+      &[0] => {
+        self
+          .handle_login::<_, ClientLoginFlow>(socket, identifiers)
+          .await
+      }
+      // Passkey auth
+      &[1] => handle_passkey_login(socket, passkey).await,
+      other => Err(anyhow!(
+        "Receieved invalid login type pattern: {other:?}"
+      )),
+    }
+  }
+}
+
+async fn handle_passkey_login(
+  socket: &mut TungsteniteWebsocket,
+  // for legacy auth
+  passkey: &str,
+) -> anyhow::Result<()> {
+  let res = async {
+    let mut passkey = if passkey.is_empty() {
+      core_config()
+        .passkey
+        .as_deref()
+        .context("Periphery requires passkey auth")?
+        .as_bytes()
+        .to_vec()
+    } else {
+      passkey.as_bytes().to_vec()
+    };
+    passkey.push(MessageState::Successful.as_byte());
+
+    socket
+      .send(passkey.into())
+      .await
+      .context("Failed to send passkey")?;
+
+    // Receive login state message and return based on value
+    let state_msg = socket
+      .recv_bytes()
+      .await
+      .context("Failed to receive authentication state message")?;
+    let state = state_msg.last().context(
+      "Authentication state message did not contain state byte",
+    )?;
+    match MessageState::from_byte(*state) {
+      MessageState::Successful => anyhow::Ok(()),
+      _ => Err(deserialize_error_bytes(
+        &state_msg[..(state_msg.len() - 1)],
+      )),
+    }
+  }
+  .await;
+  if let Err(e) = res {
+    let mut bytes = serialize_error_bytes(&e);
+    bytes.push(MessageState::Failed.as_byte());
+    if let Err(e) = socket
+      .send(bytes.into())
+      .await
+      .context("Failed to send login failed to client")
+    {
+      // Log additional error
+      warn!("{e:#}");
+    }
+    // Close socket
+    let _ = socket.close(None).await;
+    // Return the original error
+    Err(e)
+  } else {
+    Ok(())
+  }
+}

bin/core/src/connection/mod.rs

@@ -0,0 +1,480 @@
+use std::{
+  sync::{
+    Arc,
+    atomic::{self, AtomicBool},
+  },
+  time::Duration,
+};
+
+use anyhow::anyhow;
+use bytes::Bytes;
+use cache::CloneCache;
+use database::mungos::{by_id::update_one_by_id, mongodb::bson::doc};
+use komodo_client::entities::{
+  builder::{AwsBuilderConfig, UrlBuilderConfig},
+  optional_str,
+  server::Server,
+};
+use serror::serror_into_anyhow_error;
+use tokio::sync::{
+  RwLock,
+  mpsc::{Sender, error::SendError},
+};
+use tokio_util::sync::CancellationToken;
+use transport::{
+  auth::{
+    ConnectionIdentifiers, LoginFlow, LoginFlowArgs,
+    PublicKeyValidator,
+  },
+  bytes::id_from_transport_bytes,
+  channel::{BufferedReceiver, buffered_channel},
+  websocket::{
+    Websocket, WebsocketMessage, WebsocketReceiver as _,
+    WebsocketSender as _,
+  },
+};
+
+use crate::{
+  config::{core_private_key, periphery_public_keys},
+  periphery::ConnectionChannels,
+  state::db_client,
+};
+
+pub mod client;
+pub mod server;
+
+#[derive(Default)]
+pub struct PeripheryConnections(
+  CloneCache<String, Arc<PeripheryConnection>>,
+);
+
+impl PeripheryConnections {
+  /// Insert a recreated connection.
+  /// Ensures the fields which must be persisted between
+  /// connection recreation are carried over.
+  pub async fn insert(
+    &self,
+    server_id: String,
+    args: PeripheryConnectionArgs<'_>,
+  ) -> (Arc<PeripheryConnection>, BufferedReceiver<Bytes>) {
+    let (connection, receiver) = if let Some(existing_connection) =
+      self.0.remove(&server_id).await
+    {
+      existing_connection.with_new_args(args)
+    } else {
+      PeripheryConnection::new(args)
+    };
+
+    self.0.insert(server_id, connection.clone()).await;
+
+    (connection, receiver)
+  }
+
+  pub async fn get(
+    &self,
+    server_id: &String,
+  ) -> Option<Arc<PeripheryConnection>> {
+    self.0.get(server_id).await
+  }
+
+  /// Remove and cancel connection
+  pub async fn remove(
+    &self,
+    server_id: &String,
+  ) -> Option<Arc<PeripheryConnection>> {
+    self
+      .0
+      .remove(server_id)
+      .await
+      .inspect(|connection| connection.cancel())
+  }
+}
+
+/// The configurable args of a connection
+#[derive(Debug, Clone, Copy, PartialEq)]
+pub struct PeripheryConnectionArgs<'a> {
+  /// Usually the server id
+  pub id: &'a str,
+  pub address: Option<&'a str>,
+  periphery_public_key: Option<&'a str>,
+}
+
+impl PublicKeyValidator for PeripheryConnectionArgs<'_> {
+  type ValidationResult = String;
+  async fn validate(
+    &self,
+    public_key: String,
+  ) -> anyhow::Result<Self::ValidationResult> {
+    let invalid_error = || {
+      spawn_update_attempted_public_key(
+        self.id.to_string(),
+        Some(public_key.clone()),
+      );
+      let e = anyhow!("{public_key} is invalid")
+        .context(
+          "Ensure public key matches configured Periphery Public Key",
+        )
+        .context("Core failed to validate Periphery public key");
+      e
+    };
+    let core_to_periphery = self.address.is_some();
+    match (self.periphery_public_key, core_to_periphery) {
+      // The key matches expected.
+      (Some(expected), _) if public_key == expected => Ok(public_key),
+      // Explicit auth failed.
+      (Some(_), _) => Err(invalid_error()),
+      // Core -> Periphery connections with no explicit
+      // Periphery public key are not validated.
+      (None, true) => Ok(public_key),
+      // Periphery -> Core connections with no explicit
+      // Periphery public key can fall back to Core config `periphery_public_keys` if defined.
+      (None, false) => {
+        let expected =
+          periphery_public_keys().ok_or_else(invalid_error)?;
+        if expected
+          .iter()
+          .any(|expected| public_key == expected.as_str())
+        {
+          Ok(public_key)
+        } else {
+          Err(invalid_error())
+        }
+      }
+    }
+  }
+}
+
+impl<'a> PeripheryConnectionArgs<'a> {
+  pub fn from_server(server: &'a Server) -> Self {
+    Self {
+      id: &server.id,
+      address: optional_str(&server.config.address),
+      periphery_public_key: optional_str(&server.info.public_key),
+    }
+  }
+
+  pub fn from_url_builder(
+    id: &'a str,
+    config: &'a UrlBuilderConfig,
+  ) -> Self {
+    Self {
+      id,
+      address: optional_str(&config.address),
+      periphery_public_key: optional_str(
+        &config.periphery_public_key,
+      ),
+    }
+  }
+
+  pub fn from_aws_builder(
+    id: &'a str,
+    address: &'a str,
+    config: &'a AwsBuilderConfig,
+  ) -> Self {
+    Self {
+      id,
+      address: Some(address),
+      periphery_public_key: optional_str(
+        &config.periphery_public_key,
+      ),
+    }
+  }
+
+  pub fn to_owned(self) -> OwnedPeripheryConnectionArgs {
+    OwnedPeripheryConnectionArgs {
+      id: self.id.to_string(),
+      address: self.address.map(str::to_string),
+      periphery_public_key: self
+        .periphery_public_key
+        .map(str::to_string),
+    }
+  }
+
+  pub fn matches<'b>(
+    self,
+    args: impl Into<PeripheryConnectionArgs<'b>>,
+  ) -> bool {
+    self == args.into()
+  }
+}
+
+#[derive(Debug, Clone)]
+pub struct OwnedPeripheryConnectionArgs {
+  /// Usually the Server id.
+  pub id: String,
+  /// Specify outbound connection address.
+  /// Inbound connections have this as None
+  pub address: Option<String>,
+  /// The public key to expect Periphery to have.
+  /// If None, must have 'periphery_public_keys' set
+  /// in Core config, or will error
+  pub periphery_public_key: Option<String>,
+}
+
+impl OwnedPeripheryConnectionArgs {
+  pub fn borrow(&self) -> PeripheryConnectionArgs<'_> {
+    PeripheryConnectionArgs {
+      id: &self.id,
+      address: self.address.as_deref(),
+      periphery_public_key: self.periphery_public_key.as_deref(),
+    }
+  }
+}
+
+impl From<PeripheryConnectionArgs<'_>>
+  for OwnedPeripheryConnectionArgs
+{
+  fn from(value: PeripheryConnectionArgs<'_>) -> Self {
+    value.to_owned()
+  }
+}
+
+impl<'a> From<&'a OwnedPeripheryConnectionArgs>
+  for PeripheryConnectionArgs<'a>
+{
+  fn from(value: &'a OwnedPeripheryConnectionArgs) -> Self {
+    value.borrow()
+  }
+}
+
+#[derive(Debug)]
+pub struct PeripheryConnection {
+  /// The connection args
+  pub args: OwnedPeripheryConnectionArgs,
+  /// Send and receive bytes over the connection socket.
+  pub sender: Sender<Bytes>,
+  /// Cancel the connection
+  pub cancel: CancellationToken,
+  /// Whether Periphery is currently connected.
+  pub connected: AtomicBool,
+  // These fields must be maintained if new connection replaces old
+  // at the same server id.
+  /// Stores latest connection error
+  pub error: Arc<RwLock<Option<serror::Serror>>>,
+  /// Forward bytes from Periphery to specific channel handlers.
+  pub channels: Arc<ConnectionChannels>,
+}
+
+impl PeripheryConnection {
+  pub fn new(
+    args: impl Into<OwnedPeripheryConnectionArgs>,
+  ) -> (Arc<PeripheryConnection>, BufferedReceiver<Bytes>) {
+    let (sender, receiever) = buffered_channel();
+    (
+      PeripheryConnection {
+        sender,
+        args: args.into(),
+        cancel: CancellationToken::new(),
+        connected: AtomicBool::new(false),
+        error: Default::default(),
+        channels: Default::default(),
+      }
+      .into(),
+      receiever,
+    )
+  }
+
+  pub fn with_new_args(
+    &self,
+    args: impl Into<OwnedPeripheryConnectionArgs>,
+  ) -> (Arc<PeripheryConnection>, BufferedReceiver<Bytes>) {
+    // Ensure this connection is cancelled.
+    self.cancel();
+    let (sender, receiever) = buffered_channel();
+    (
+      PeripheryConnection {
+        sender,
+        args: args.into(),
+        cancel: CancellationToken::new(),
+        connected: AtomicBool::new(false),
+        error: self.error.clone(),
+        channels: self.channels.clone(),
+      }
+      .into(),
+      receiever,
+    )
+  }
+
+  pub async fn handle_login<W: Websocket, L: LoginFlow>(
+    &self,
+    socket: &mut W,
+    identifiers: ConnectionIdentifiers<'_>,
+  ) -> anyhow::Result<()> {
+    L::login(LoginFlowArgs {
+      socket,
+      identifiers,
+      private_key: core_private_key(),
+      public_key_validator: self.args.borrow(),
+    })
+    .await?;
+    // Clear attempted public key after successful login
+    spawn_update_attempted_public_key(self.args.id.clone(), None);
+    Ok(())
+  }
+
+  pub async fn handle_socket<W: Websocket>(
+    &self,
+    socket: W,
+    receiver: &mut BufferedReceiver<Bytes>,
+  ) {
+    let cancel = self.cancel.child_token();
+
+    self.set_connected(true);
+    self.clear_error().await;
+
+    let (mut ws_write, mut ws_read) = socket.split();
+
+    let forward_writes = async {
+      loop {
+        let next = tokio::select! {
+          next = receiver.recv() => next,
+          _ = cancel.cancelled() => break,
+        };
+
+        let message = match next {
+          Some(request) => Bytes::copy_from_slice(request),
+          // Sender Dropped (shouldn't happen, a reference is held on 'connection').
+          None => break,
+        };
+
+        match ws_write.send(message).await {
+          Ok(_) => receiver.clear_buffer(),
+          Err(e) => {
+            self.set_error(e.into()).await;
+            break;
+          }
+        }
+      }
+      // Cancel again if not already
+      let _ = ws_write.close(None).await;
+      cancel.cancel();
+    };
+
+    let handle_reads = async {
+      loop {
+        let next = tokio::select! {
+          next = ws_read.recv() => next,
+          _ = cancel.cancelled() => break,
+        };
+
+        match next {
+          Ok(WebsocketMessage::Binary(bytes)) => {
+            self.handle_incoming_bytes(bytes).await
+          }
+          Ok(WebsocketMessage::Close(_))
+          | Ok(WebsocketMessage::Closed) => {
+            self.set_error(anyhow!("Connection closed")).await;
+            break;
+          }
+          Err(e) => {
+            self.set_error(e.into()).await;
+          }
+        };
+      }
+      // Cancel again if not already
+      cancel.cancel();
+    };
+
+    tokio::join!(forward_writes, handle_reads);
+
+    self.set_connected(false);
+  }
+
+  pub async fn handle_incoming_bytes(&self, bytes: Bytes) {
+    let id = match id_from_transport_bytes(&bytes) {
+      Ok(res) => res,
+      Err(e) => {
+        // TODO: handle better
+        warn!("Failed to read id | {e:#}");
+        return;
+      }
+    };
+    let Some(channel) = self.channels.get(&id).await else {
+      // TODO: handle better
+      debug!("Failed to send response | No response channel found");
+      return;
+    };
+    if let Err(e) = channel.send(bytes).await {
+      // TODO: handle better
+      warn!("Failed to send response | Channel failure | {e:#}");
+    }
+  }
+
+  pub async fn send(
+    &self,
+    value: Bytes,
+  ) -> Result<(), SendError<Bytes>> {
+    self.sender.send(value).await
+  }
+
+  pub fn set_connected(&self, connected: bool) {
+    self.connected.store(connected, atomic::Ordering::Relaxed);
+  }
+
+  pub fn connected(&self) -> bool {
+    self.connected.load(atomic::Ordering::Relaxed)
+  }
+
+  /// Polls connected 3 times (500ms in between) before bailing.
+  pub async fn bail_if_not_connected(&self) -> anyhow::Result<()> {
+    const POLL_TIMES: usize = 3;
+    for i in 0..POLL_TIMES {
+      if self.connected() {
+        return Ok(());
+      }
+      if i < POLL_TIMES - 1 {
+        tokio::time::sleep(Duration::from_millis(500)).await;
+      }
+    }
+    if let Some(e) = self.error().await {
+      Err(serror_into_anyhow_error(e))
+    } else {
+      Err(anyhow!("Server is not currently connected"))
+    }
+  }
+
+  pub async fn error(&self) -> Option<serror::Serror> {
+    self.error.read().await.clone()
+  }
+
+  pub async fn set_error(&self, e: anyhow::Error) {
+    let mut error = self.error.write().await;
+    *error = Some(e.into());
+  }
+
+  pub async fn clear_error(&self) {
+    let mut error = self.error.write().await;
+    *error = None;
+  }
+
+  pub fn cancel(&self) {
+    self.cancel.cancel();
+  }
+}
+
+/// Spawn task to set the 'attempted_public_key'
+/// for easy manual connection acceptance later on.
+fn spawn_update_attempted_public_key(
+  id: String,
+  public_key: impl Into<Option<String>>,
+) {
+  let public_key = public_key.into();
+  tokio::spawn(async move {
+    if let Err(e) = update_one_by_id(
+      &db_client().servers,
+      &id,
+      doc! {
+        "$set": {
+          "info.attempted_public_key": &public_key.as_deref().unwrap_or_default(),
+        }
+      },
+      None,
+    )
+    .await
+    {
+      warn!(
+        "Failed to update attempted public_key for Server {id} | {e:?}"
+      );
+    };
+  });
+}

bin/core/src/connection/server.rs

@@ -0,0 +1,349 @@
+use std::{str::FromStr, time::Duration};
+
+use anyhow::{Context, anyhow};
+use axum::{
+  extract::{Query, WebSocketUpgrade},
+  http::{HeaderMap, StatusCode},
+  response::Response,
+};
+use bytes::Bytes;
+use database::mungos::mongodb::bson::{doc, oid::ObjectId};
+use komodo_client::{
+  api::write::{CreateBuilder, CreateServer, UpdateResourceMeta},
+  entities::{
+    builder::{PartialBuilderConfig, PartialServerBuilderConfig},
+    onboarding_key::OnboardingKey,
+    server::{PartialServerConfig, Server},
+    user::system_user,
+  },
+};
+use resolver_api::Resolve;
+use serror::{AddStatusCode, AddStatusCodeError};
+use transport::{
+  MessageState, PeripheryConnectionQuery,
+  auth::{
+    HeaderConnectionIdentifiers, LoginFlow, LoginFlowArgs,
+    PublicKeyValidator, ServerLoginFlow,
+  },
+  websocket::{Websocket, axum::AxumWebsocket},
+};
+
+use crate::{
+  api::write::WriteArgs,
+  config::core_private_key,
+  helpers::query::id_or_name_filter,
+  resource::KomodoResource,
+  state::{db_client, periphery_connections},
+};
+
+use super::PeripheryConnectionArgs;
+
+pub async fn handler(
+  Query(PeripheryConnectionQuery {
+    server: server_query,
+  }): Query<PeripheryConnectionQuery>,
+  mut headers: HeaderMap,
+  ws: WebSocketUpgrade,
+) -> serror::Result<Response> {
+  let identifiers =
+    HeaderConnectionIdentifiers::extract(&mut headers)
+      .status_code(StatusCode::UNAUTHORIZED)?;
+
+  if server_query.is_empty() {
+    return Err(
+      anyhow!("Must provide non-empty server specifier")
+        .status_code(StatusCode::UNAUTHORIZED),
+    );
+  }
+
+  // Handle connection vs. onboarding flow.
+  match Server::coll()
+    .find_one(id_or_name_filter(&server_query))
+    .await
+    .context("Failed to query database for Server")?
+  {
+    Some(server) => {
+      existing_server_handler(server_query, server, identifiers, ws)
+        .await
+    }
+    None if ObjectId::from_str(&server_query).is_err() => {
+      onboard_server_handler(server_query, identifiers, ws).await
+    }
+    None => Err(
+      anyhow!("Must provide name based Server specifier for onboarding flow, name cannot be valid ObjectId (hex)")
+        .status_code(StatusCode::UNAUTHORIZED),
+    ),
+  }
+}
+
+async fn existing_server_handler(
+  server_query: String,
+  server: Server,
+  identifiers: HeaderConnectionIdentifiers,
+  ws: WebSocketUpgrade,
+) -> serror::Result<Response> {
+  if !server.config.enabled {
+    return Err(anyhow!("Server is Disabled."))
+      .status_code(StatusCode::BAD_REQUEST);
+  }
+
+  if !server.config.address.is_empty() {
+    return Err(anyhow!(
+      "Server is configured to use a Core -> Periphery connection."
+    ))
+    .status_code(StatusCode::BAD_REQUEST);
+  }
+
+  let connections = periphery_connections();
+
+  // Ensure connected server can't get bumped off the connection.
+  // Treat this as authorization issue.
+  if let Some(existing_connection) = connections.get(&server.id).await
+    && existing_connection.connected()
+  {
+    return Err(
+      anyhow!("A Server '{server_query}' is already connected")
+        .status_code(StatusCode::UNAUTHORIZED),
+    );
+  }
+
+  let (connection, mut receiver) = periphery_connections()
+    .insert(
+      server.id.clone(),
+      PeripheryConnectionArgs::from_server(&server),
+    )
+    .await;
+
+  Ok(ws.on_upgrade(|socket| async move {
+    let query =
+      format!("server={}", urlencoding::encode(&server_query));
+    let mut socket = AxumWebsocket(socket);
+
+    if let Err(e) = socket.send(Bytes::from_owner([0])).await.context(
+      "Failed to send the login flow indicator over connnection",
+    ) {
+      connection.set_error(e).await;
+      return;
+    };
+
+    if let Err(e) = connection
+      .handle_login::<_, ServerLoginFlow>(
+        &mut socket,
+        identifiers.build(query.as_bytes()),
+      )
+      .await
+    {
+      connection.set_error(e).await;
+      return;
+    }
+
+    connection.handle_socket(socket, &mut receiver).await
+  }))
+}
+
+async fn onboard_server_handler(
+  server_query: String,
+  identifiers: HeaderConnectionIdentifiers,
+  ws: WebSocketUpgrade,
+) -> serror::Result<Response> {
+  Ok(ws.on_upgrade(|socket| async move {
+    let query =
+      format!("server={}", urlencoding::encode(&server_query));
+    let mut socket = AxumWebsocket(socket);
+
+    if let Err(e) = socket.send(Bytes::from_owner([1])).await.context(
+      "Failed to send the login flow indicator over connnection",
+    ).context("Server onboarding error") {
+      warn!("{e:#}");
+      return;
+    };
+
+    let onboarding_key = match ServerLoginFlow::login(LoginFlowArgs {
+      socket: &mut socket,
+      identifiers: identifiers.build(query.as_bytes()),
+      private_key: core_private_key(),
+      public_key_validator: CreationKeyValidator,
+    })
+    .await
+    {
+      Ok(onboarding_key) => onboarding_key,
+      Err(e) => {
+        debug!("Server {server_query} failed to onboard | {e:#}");
+        return;
+      }
+    };
+
+    let res = socket
+      .recv_bytes()
+      .with_timeout(Duration::from_secs(2))
+      .await
+      .and_then(|res| {
+        res.and_then(|public_key_bytes| {
+          String::from_utf8(public_key_bytes.into())
+            .context("Public key bytes are not valid utf8")
+        })
+      });
+
+    // Post onboarding login 1: Receive public key
+    let public_key = match res
+    {
+      Ok(public_key) => public_key,
+      Err(e) => {
+        warn!("Server {server_query} failed to onboard | failed to receive Server public key | {e:#}");
+        return;
+      }
+    };
+
+    
+
+    let server_id = match create_server_maybe_builder(
+      server_query,
+      public_key,
+      onboarding_key.copy_server,
+      onboarding_key.tags,
+      onboarding_key.create_builder
+    ).await {
+      Ok(server_id) => server_id,
+      Err(e) => {
+        warn!("{e:#}");
+        if let Err(e) = socket
+          .send_error(&e)
+          .await
+          .context("Failed to send Server creation failed to client")
+        {
+          // Log additional error
+          warn!("{e:#}");
+        }
+        return;
+      }
+    };
+
+    if let Err(e) = socket
+      .send(MessageState::Successful.into())
+      .await
+      .context("Failed to send Server creation successful to client")
+    {
+      // Log additional error
+      warn!("{e:#}");
+    }
+
+    // Server created, close and trigger reconnect
+    // and handling using existing server handler.
+    let _ = socket.close(None).await;
+
+    // Add the server to onboarding key "Onboarded"
+    let res = db_client()
+      .onboarding_keys
+      .update_one(
+        doc! { "public_key": &onboarding_key.public_key },
+        doc! { "$push": { "onboarded": server_id } },
+      ).await;
+    if let Err(e) = res {
+      warn!("Failed to update onboarding key 'onboarded' | {e:?}");
+    }
+  }))
+}
+
+async fn create_server_maybe_builder(
+  server_query: String,
+  public_key: String,
+  copy_server: String,
+  tags: Vec<String>,
+  create_builder: bool,
+) -> anyhow::Result<String> {
+  let config = if copy_server.is_empty() {
+    PartialServerConfig {
+      enabled: Some(true),
+      ..Default::default()
+    }
+  } else {
+    let config = match db_client().servers.find_one(id_or_name_filter(&copy_server)).await {
+      Ok(Some(server)) => server.config,
+      Ok(None) => {
+        warn!("Server onboarding: Failed to find Server {}", copy_server);
+        Default::default()
+      }
+      Err(e) => {
+        warn!("Failed to query database for onboarding key 'copy_server' | {e:?}");
+        Default::default()
+      }
+    };
+    PartialServerConfig {
+      enabled: Some(true),
+      address: None,
+      ..config.into()
+    }
+  };
+
+  let args = WriteArgs {
+    user: system_user().to_owned(),
+  };
+
+  let server = CreateServer {
+    name: server_query.clone(),
+    config,
+    public_key: Some(public_key),
+  }
+  .resolve(&args)
+  .await
+  .map_err(|e| e.error)
+  .context("Server onboarding flow failed at Server creation")?;
+
+  // Don't need to fail, only warn on this
+  if let Err(e) = (UpdateResourceMeta {
+    target: (&server).into(),
+    tags: Some(tags),
+    description: None,
+    template: None,
+  })
+  .resolve(&args)
+  .await
+  .map_err(|e| e.error)
+  .context("Server onboarding flow failed at Server creation")
+  {
+    warn!("{e:#}");
+  };
+
+  if create_builder {
+    // Don't need to fail, only warn on this
+    if let Err(e) = (CreateBuilder {
+      name: server_query,
+      config: PartialBuilderConfig::Server(
+        PartialServerBuilderConfig {
+          server_id: Some(server.id.clone()),
+        },
+      ),
+    })
+    .resolve(&args)
+    .await
+    .map_err(|e| e.error)
+    .context("Server onboarding flow failed at Builder creation")
+    {
+      warn!("{e:#}");
+    };
+  }
+
+  Ok(server.id)
+}
+
+struct CreationKeyValidator;
+
+impl PublicKeyValidator for CreationKeyValidator {
+  type ValidationResult = OnboardingKey;
+  async fn validate(
+    &self,
+    public_key: String,
+  ) -> anyhow::Result<Self::ValidationResult> {
+    let onboarding_key = db_client()
+      .onboarding_keys
+      .find_one(doc! { "public_key": &public_key })
+      .await
+      .context("Failed to query database for Server onboarding keys")?
+      .context("Matching Server onboarding key not found")?;
+    if onboarding_key.enabled {
+      Ok(onboarding_key)
+    } else {
+      Err(anyhow!("Onboarding key is disabled"))
+    }
+  }
+}

bin/core/src/helpers/action_state.rs

@@ -1,6 +1,7 @@
 use std::sync::{Arc, Mutex};
 
 use anyhow::anyhow;
+use cache::CloneCache;
 use komodo_client::{
   busy::Busy,
   entities::{
@@ -12,20 +13,19 @@ use komodo_client::{
   },
 };
 
-use super::cache::Cache;
-
 #[derive(Default)]
 pub struct ActionStates {
-  pub server: Cache<String, Arc<ActionState<ServerActionState>>>,
-  pub stack: Cache<String, Arc<ActionState<StackActionState>>>,
+  pub server: CloneCache<String, Arc<ActionState<ServerActionState>>>,
+  pub stack: CloneCache<String, Arc<ActionState<StackActionState>>>,
   pub deployment:
-    Cache<String, Arc<ActionState<DeploymentActionState>>>,
-  pub build: Cache<String, Arc<ActionState<BuildActionState>>>,
-  pub repo: Cache<String, Arc<ActionState<RepoActionState>>>,
+    CloneCache<String, Arc<ActionState<DeploymentActionState>>>,
+  pub build: CloneCache<String, Arc<ActionState<BuildActionState>>>,
+  pub repo: CloneCache<String, Arc<ActionState<RepoActionState>>>,
   pub procedure:
-    Cache<String, Arc<ActionState<ProcedureActionState>>>,
-  pub action: Cache<String, Arc<ActionState<ActionActionState>>>,
-  pub sync: Cache<String, Arc<ActionState<ResourceSyncActionState>>>,
+    CloneCache<String, Arc<ActionState<ProcedureActionState>>>,
+  pub action: CloneCache<String, Arc<ActionState<ActionActionState>>>,
+  pub sync:
+    CloneCache<String, Arc<ActionState<ResourceSyncActionState>>>,
 }
 
 /// Need to be able to check "busy" with write lock acquired.

bin/core/src/helpers/builder.rs

@@ -1,6 +1,7 @@
 use std::time::Duration;
 
 use anyhow::{Context, anyhow};
+use database::mungos::mongodb::bson::oid::ObjectId;
 use formatting::muted;
 use komodo_client::entities::{
   Version,
@@ -9,10 +10,7 @@ use komodo_client::entities::{
   server::Server,
   update::{Log, Update},
 };
-use periphery_client::{
-  PeripheryClient,
-  api::{self, GetVersionResponse},
-};
+use periphery_client::api::{self, GetVersionResponse};
 
 use crate::{
   cloud::{
@@ -22,8 +20,9 @@ use crate::{
       terminate_ec2_instance_with_retry,
     },
   },
-  config::core_config,
+  connection::PeripheryConnectionArgs,
   helpers::update::update_update,
+  periphery::PeripheryClient,
   resource,
 };
 
@@ -47,27 +46,29 @@ pub async fn get_builder_periphery(
           "Builder has not yet configured an address"
         ));
       }
+      // TODO: Dont use builder id, or will be problems
+      // with simultaneous spawned builders.
       let periphery = PeripheryClient::new(
-        config.address,
-        if config.passkey.is_empty() {
-          core_config().passkey.clone()
-        } else {
-          config.passkey
-        },
-        Duration::from_secs(3),
-      );
+        PeripheryConnectionArgs::from_url_builder(
+          &ObjectId::new().to_hex(),
+          &config,
+        ),
+        config.insecure_tls,
+        &config.passkey,
+      )
+      .await?;
       periphery
         .health_check()
         .await
         .context("Url Builder failed health check")?;
-      Ok((periphery, BuildCleanupData::Server))
+      Ok((periphery, BuildCleanupData::Url))
     }
     BuilderConfig::Server(config) => {
       if config.server_id.is_empty() {
         return Err(anyhow!("Builder has not configured a server"));
       }
       let server = resource::get::<Server>(&config.server_id).await?;
-      let periphery = periphery_client(&server)?;
+      let periphery = periphery_client(&server).await?;
       Ok((periphery, BuildCleanupData::Server))
     }
     BuilderConfig::Aws(config) => {
@@ -90,10 +91,8 @@ async fn get_aws_builder(
   let Ec2Instance { instance_id, ip } =
     launch_ec2_instance(&instance_name, &config).await?;
 
-  info!("ec2 instance launched");
-
   let log = Log {
-    stage: "start build instance".to_string(),
+    stage: "Start Build Instance".to_string(),
     success: true,
     stdout: start_aws_builder_log(&instance_id, &ip, &config),
     start_ts: start_create_ts,
@@ -105,14 +104,21 @@ async fn get_aws_builder(
 
   update_update(update.clone()).await?;
 
-  let protocol = if config.use_https { "https" } else { "http" };
+  let protocol = if config.use_https { "wss" } else { "ws" };
+
+  // TODO: Handle ad-hoc (non server) periphery connections. These don't have ids.
   let periphery_address =
     format!("{protocol}://{ip}:{}", config.port);
   let periphery = PeripheryClient::new(
-    &periphery_address,
-    &core_config().passkey,
-    Duration::from_secs(3),
-  );
+    PeripheryConnectionArgs::from_aws_builder(
+      &ObjectId::new().to_hex(),
+      &periphery_address,
+      &config,
+    ),
+    config.insecure_tls,
+    "",
+  )
+  .await?;
 
   let start_connect_ts = komodo_timestamp();
   let mut res = Ok(GetVersionResponse {
@@ -166,6 +172,7 @@ async fn get_aws_builder(
 
 #[instrument(skip(update))]
 pub async fn cleanup_builder_instance(
+  periphery: PeripheryClient,
   cleanup_data: BuildCleanupData,
   update: &mut Update,
 ) {
@@ -173,10 +180,14 @@ pub async fn cleanup_builder_instance(
     BuildCleanupData::Server => {
       // Nothing to clean up
     }
+    BuildCleanupData::Url => {
+      periphery.cleanup().await;
+    }
     BuildCleanupData::Aws {
       instance_id,
       region,
     } => {
+      periphery.cleanup().await;
       let _instance_id = instance_id.clone();
       tokio::spawn(async move {
         let _ =

bin/core/src/helpers/cache.rs

@@ -1,83 +0,0 @@
-use std::{collections::HashMap, hash::Hash};
-
-use tokio::sync::RwLock;
-
-#[derive(Default)]
-pub struct Cache<K: PartialEq + Eq + Hash, T: Clone + Default> {
-  cache: RwLock<HashMap<K, T>>,
-}
-
-impl<
-  K: PartialEq + Eq + Hash + std::fmt::Debug + Clone,
-  T: Clone + Default,
-> Cache<K, T>
-{
-  #[instrument(level = "debug", skip(self))]
-  pub async fn get(&self, key: &K) -> Option<T> {
-    self.cache.read().await.get(key).cloned()
-  }
-
-  #[instrument(level = "debug", skip(self))]
-  pub async fn get_or_insert_default(&self, key: &K) -> T {
-    let mut lock = self.cache.write().await;
-    match lock.get(key).cloned() {
-      Some(item) => item,
-      None => {
-        let item: T = Default::default();
-        lock.insert(key.clone(), item.clone());
-        item
-      }
-    }
-  }
-
-  #[instrument(level = "debug", skip(self))]
-  pub async fn get_list(&self) -> Vec<T> {
-    let cache = self.cache.read().await;
-    cache.values().cloned().collect()
-  }
-
-  #[instrument(level = "debug", skip(self))]
-  pub async fn insert<Key>(&self, key: Key, val: T)
-  where
-    T: std::fmt::Debug,
-    Key: Into<K> + std::fmt::Debug,
-  {
-    self.cache.write().await.insert(key.into(), val);
-  }
-
-  // #[instrument(level = "debug", skip(self, handler))]
-  // pub async fn update_entry<Key>(
-  //   &self,
-  //   key: Key,
-  //   handler: impl Fn(&mut T),
-  // ) where
-  //   Key: Into<K> + std::fmt::Debug,
-  // {
-  //   let mut cache = self.cache.write().await;
-  //   handler(cache.entry(key.into()).or_default());
-  // }
-
-  // #[instrument(level = "debug", skip(self))]
-  // pub async fn clear(&self) {
-  //   self.cache.write().await.clear();
-  // }
-
-  #[instrument(level = "debug", skip(self))]
-  pub async fn remove(&self, key: &K) {
-    self.cache.write().await.remove(key);
-  }
-}
-
-// impl<
-//   K: PartialEq + Eq + Hash + std::fmt::Debug + Clone,
-//   T: Clone + Default + Busy,
-// > Cache<K, T>
-// {
-//   #[instrument(level = "debug", skip(self))]
-//   pub async fn busy(&self, id: &K) -> bool {
-//     match self.get(id).await {
-//       Some(state) => state.busy(),
-//       None => false,
-//     }
-//   }
-// }

bin/core/src/helpers/mod.rs

@@ -1,4 +1,4 @@
-use std::{fmt::Write, time::Duration};
+use std::fmt::Write;
 
 use anyhow::{Context, anyhow};
 use database::mongo_indexed::Document;
@@ -15,15 +15,16 @@ use komodo_client::entities::{
   stack::Stack,
   user::User,
 };
-use periphery_client::PeripheryClient;
 use rand::Rng;
 
-use crate::{config::core_config, state::db_client};
+use crate::{
+  config::core_config, connection::PeripheryConnectionArgs,
+  periphery::PeripheryClient, state::db_client,
+};
 
 pub mod action_state;
 pub mod all_resources;
 pub mod builder;
-pub mod cache;
 pub mod channel;
 pub mod maintenance;
 pub mod matcher;
@@ -185,24 +186,18 @@ pub async fn registry_token(
 
 //
 
-pub fn periphery_client(
+pub async fn periphery_client(
   server: &Server,
 ) -> anyhow::Result<PeripheryClient> {
   if !server.config.enabled {
     return Err(anyhow!("server not enabled"));
   }
-
-  let client = PeripheryClient::new(
-    &server.config.address,
-    if server.config.passkey.is_empty() {
-      &core_config().passkey
-    } else {
-      &server.config.passkey
-    },
-    Duration::from_secs(server.config.timeout_seconds as u64),
-  );
-
-  Ok(client)
+  PeripheryClient::new(
+    PeripheryConnectionArgs::from_server(server),
+    server.config.insecure_tls,
+    &server.config.passkey,
+  )
+  .await
 }
 
 #[instrument]

bin/core/src/helpers/procedure.rs

@@ -1209,6 +1209,23 @@ async fn execute_execution(
       )
       .await?
     }
+    Execution::RotateAllServerKeys(req) => {
+      let req = ExecuteRequest::RotateAllServerKeys(req);
+      let update = init_execution_update(&req, &user).await?;
+      let ExecuteRequest::RotateAllServerKeys(req) = req else {
+        unreachable!()
+      };
+      let update_id = update.id.clone();
+      handle_resolve_result(
+        req
+          .resolve(&ExecuteArgs { user, update })
+          .await
+          .map_err(|e| e.error)
+          .context("Failed at RotateAllServerKeys"),
+        &update_id,
+      )
+      .await?
+    }
     Execution::Sleep(req) => {
       let duration = Duration::from_millis(req.duration_ms as u64);
       tokio::time::sleep(duration).await;

bin/core/src/helpers/prune.rs

@@ -4,7 +4,7 @@ use async_timing_util::{
 };
 use database::mungos::{find::find_collect, mongodb::bson::doc};
 use futures::{StreamExt, stream::FuturesUnordered};
-use periphery_client::api::image::PruneImages;
+use periphery_client::api::docker::PruneImages;
 
 use crate::{config::core_config, state::db_client};
 
@@ -41,7 +41,10 @@ async fn prune_images() -> anyhow::Result<()> {
   .map(|server| async move {
     (
       async {
-        periphery_client(&server)?.request(PruneImages {}).await
+        periphery_client(&server)
+          .await?
+          .request(PruneImages {})
+          .await
       }
       .await,
       server,
@@ -51,8 +54,8 @@ async fn prune_images() -> anyhow::Result<()> {
 
   while let Some((res, server)) = futures.next().await {
     if let Err(e) = res {
-      error!(
-        "failed to prune images on server {} ({}) | {e:#}",
+      warn!(
+        "failed to prune images on Server {} ({}) | {e:#}",
         server.name, server.id
       )
     }

bin/core/src/helpers/query.rs

@@ -1,11 +1,6 @@
-use std::{
-  collections::HashMap,
-  str::FromStr,
-  sync::{Arc, OnceLock},
-};
+use std::{collections::HashMap, str::FromStr};
 
 use anyhow::{Context, anyhow};
-use async_timing_util::{ONE_MIN_MS, unix_timestamp_ms};
 use database::mungos::{
   find::find_collect,
   mongodb::{
@@ -30,7 +25,6 @@ use komodo_client::{
     repo::Repo,
     server::{Server, ServerState},
     stack::{Stack, StackServiceNames, StackState},
-    stats::SystemInformation,
     sync::ResourceSync,
     tag::Tag,
     update::Update,
@@ -39,8 +33,6 @@ use komodo_client::{
     variable::Variable,
   },
 };
-use periphery_client::api::stats;
-use tokio::sync::Mutex;
 
 use crate::{
   config::core_config,
@@ -54,8 +46,6 @@ use crate::{
   },
 };
 
-use super::periphery_client;
-
 // user: Id or username
 #[instrument(level = "debug")]
 pub async fn get_user(user: &str) -> anyhow::Result<User> {
@@ -84,10 +74,11 @@ pub async fn get_server_state(server: &Server) -> ServerState {
   if !server.config.enabled {
     return ServerState::Disabled;
   }
-  // Unwrap ok: Server disabled check above
-  match super::periphery_client(server)
-    .unwrap()
-    .request(periphery_client::api::GetHealth {})
+  let Ok(periphery) = super::periphery_client(server).await else {
+    return ServerState::NotOk;
+  };
+  match periphery
+    .request(periphery_client::api::GetVersion {})
     .await
   {
     Ok(_) => ServerState::Ok,
@@ -413,39 +404,6 @@ pub async fn get_variables_and_secrets()
   Ok(VariablesAndSecrets { variables, secrets })
 }
 
-// This protects the peripheries from spam requests
-const SYSTEM_INFO_EXPIRY: u128 = ONE_MIN_MS;
-type SystemInfoCache =
-  Mutex<HashMap<String, Arc<(SystemInformation, u128)>>>;
-fn system_info_cache() -> &'static SystemInfoCache {
-  static SYSTEM_INFO_CACHE: OnceLock<SystemInfoCache> =
-    OnceLock::new();
-  SYSTEM_INFO_CACHE.get_or_init(Default::default)
-}
-
-pub async fn get_system_info(
-  server: &Server,
-) -> anyhow::Result<SystemInformation> {
-  let mut lock = system_info_cache().lock().await;
-  let res = match lock.get(&server.id) {
-    Some(cached) if cached.1 > unix_timestamp_ms() => {
-      cached.0.clone()
-    }
-    _ => {
-      let stats = periphery_client(server)?
-        .request(stats::GetSystemInformation {})
-        .await?;
-      lock.insert(
-        server.id.clone(),
-        (stats.clone(), unix_timestamp_ms() + SYSTEM_INFO_EXPIRY)
-          .into(),
-      );
-      stats
-    }
-  };
-  Ok(res)
-}
-
 /// Get last time procedure / action was run using Update query.
 /// Ignored whether run was successful.
 pub async fn get_last_run_at<R: KomodoResource>(

bin/core/src/helpers/update.rs

@@ -520,6 +520,9 @@ pub async fn init_execution_update(
     ExecuteRequest::GlobalAutoUpdate(_data) => {
       (Operation::GlobalAutoUpdate, ResourceTarget::system())
     }
+    ExecuteRequest::RotateAllServerKeys(_data) => {
+      (Operation::RotateAllServerKeys, ResourceTarget::system())
+    }
   };
 
   let mut update = make_update(target, operation, user);

bin/core/src/listener/mod.rs

@@ -2,10 +2,11 @@ use std::sync::Arc;
 
 use anyhow::anyhow;
 use axum::{Router, http::HeaderMap};
+use cache::CloneCache;
 use komodo_client::entities::resource::Resource;
 use tokio::sync::Mutex;
 
-use crate::{helpers::cache::Cache, resource::KomodoResource};
+use crate::resource::KomodoResource;
 
 mod integrations;
 mod resources;
@@ -19,7 +20,7 @@ pub fn router() -> Router {
     .nest("/gitlab", router::router::<gitlab::Gitlab>())
 }
 
-type ListenerLockCache = Cache<String, Arc<Mutex<()>>>;
+type ListenerLockCache = CloneCache<String, Arc<Mutex<()>>>;
 
 /// Implemented for all resources which can recieve webhook.
 trait CustomSecret: KomodoResource {

bin/core/src/main.rs

@@ -1,3 +1,5 @@
+#![recursion_limit = "256"]
+
 #[macro_use]
 extern crate tracing;
 
@@ -11,17 +13,19 @@ use tower_http::{
   services::{ServeDir, ServeFile},
 };
 
-use crate::config::core_config;
+use crate::config::{core_config, core_public_key};
 
 mod alert;
 mod api;
 mod auth;
 mod cloud;
 mod config;
+mod connection;
 mod helpers;
 mod listener;
 mod monitor;
 mod network;
+mod periphery;
 mod permission;
 mod resource;
 mod schedule;
@@ -36,12 +40,6 @@ async fn app() -> anyhow::Result<()> {
   dotenvy::dotenv().ok();
   let config = core_config();
   logger::init(&config.logging)?;
-  if let Err(e) =
-    rustls::crypto::aws_lc_rs::default_provider().install_default()
-  {
-    error!("Failed to install default crypto provider | {e:?}");
-    std::process::exit(1);
-  };
 
   info!("Komodo Core version: v{}", env!("CARGO_PKG_VERSION"));
 
@@ -55,6 +53,13 @@ async fn app() -> anyhow::Result<()> {
     (false, false) => info!("{:?}", config.sanitized()),
   }
 
+  // Init + log public key. Will crash if invalid private key here.
+  info!("Public Key: {}", core_public_key());
+
+  rustls::crypto::aws_lc_rs::default_provider()
+    .install_default()
+    .expect("Failed to install default crypto provider");
+
   // Init jwt client to crash on failure
   state::jwt_client();
   tokio::join!(
@@ -122,9 +127,6 @@ async fn app() -> anyhow::Result<()> {
 
   if config.ssl_enabled {
     info!("🔒 Core SSL Enabled");
-    rustls::crypto::ring::default_provider()
-      .install_default()
-      .expect("failed to install default rustls CryptoProvider");
     info!("Komodo Core starting on https://{socket_addr}");
     let ssl_config = RustlsConfig::from_pem_file(
       &config.ssl_cert_file,

bin/core/src/monitor/alert/deployment.rs

@@ -20,7 +20,7 @@ pub async fn alert_deployments(
 ) {
   let mut alerts = Vec::<Alert>::new();
   let action_states = action_states();
-  for status in deployment_status_cache().get_list().await {
+  for status in deployment_status_cache().get_values().await {
     // Don't alert if prev None
     let Some(prev) = status.prev else {
       continue;

bin/core/src/monitor/alert/server.rs

@@ -78,7 +78,7 @@ pub async fn alert_servers(
   ts: i64,
   mut servers: HashMap<String, Server>,
 ) {
-  let server_statuses = server_status_cache().get_list().await;
+  let server_statuses = server_status_cache().get_values().await;
 
   let (open_alerts, open_disk_alerts) = match get_open_alerts().await
   {
@@ -182,17 +182,24 @@ pub async fn alert_servers(
     // SERVER VERSION MISMATCH
     // ===================
     let core_version = env!("CARGO_PKG_VERSION");
-    let has_version_mismatch = server_status.state == ServerState::Ok
-      && !server_status.version.is_empty()
-      && server_status.version != "Unknown"
-      && server_status.version != core_version;
+    let mismatched_server_version =
+      if server_status.state != ServerState::Ok {
+        None
+      } else if let Some(version) =
+        server_status.periphery_info.as_ref().map(|i| &i.version)
+        && version != core_version
+      {
+        Some(version)
+      } else {
+        None
+      };
 
     let version_alert = server_alerts.as_ref().and_then(|alerts| {
       alerts.get(&AlertDataVariant::ServerVersionMismatch)
     });
 
-    match (has_version_mismatch, version_alert) {
-      (true, None) => {
+    match (mismatched_server_version, version_alert) {
+      (Some(version), None) => {
         // Only open version mismatch alert if not in maintenance and buffer is ready
         if !in_maintenance
           && buffer.ready_to_open(
@@ -211,7 +218,7 @@ pub async fn alert_servers(
               id: server_status.id.clone(),
               name: server.name.clone(),
               region: optional_string(&server.config.region),
-              server_version: server_status.version.clone(),
+              server_version: version.clone(),
               core_version: core_version.to_string(),
             },
           };
@@ -220,27 +227,27 @@ pub async fn alert_servers(
             .push((alert, server.config.send_version_mismatch_alerts))
         }
       }
-      (true, Some(alert)) => {
+      (Some(version), Some(alert)) => {
         // Update existing alert with current version info
         let mut alert = alert.clone();
         alert.data = AlertData::ServerVersionMismatch {
           id: server_status.id.clone(),
           name: server.name.clone(),
           region: optional_string(&server.config.region),
-          server_version: server_status.version.clone(),
+          server_version: version.clone(),
           core_version: core_version.to_string(),
         };
         // Don't send notification for updates
         alerts_to_update.push((alert, false));
       }
-      (false, Some(alert)) => {
+      (None, Some(alert)) => {
         // Version is now correct, close the alert
         alert_ids_to_close.push((
           alert.clone(),
           server.config.send_version_mismatch_alerts,
         ));
       }
-      (false, None) => {
+      (None, None) => {
         // Reset buffer state when no mismatch and no alert
         buffer.reset(
           server_status.id.clone(),
@@ -282,7 +289,7 @@ pub async fn alert_servers(
               name: server.name.clone(),
               region: optional_string(&server.config.region),
               percentage: server_status
-                .stats
+                .system_stats
                 .as_ref()
                 .map(|s| s.cpu_perc as f64)
                 .unwrap_or(0.0),
@@ -304,7 +311,7 @@ pub async fn alert_servers(
             name: server.name.clone(),
             region: optional_string(&server.config.region),
             percentage: server_status
-              .stats
+              .system_stats
               .as_ref()
               .map(|s| s.cpu_perc as f64)
               .unwrap_or(0.0),
@@ -320,7 +327,7 @@ pub async fn alert_servers(
           name: server.name.clone(),
           region: optional_string(&server.config.region),
           percentage: server_status
-            .stats
+            .system_stats
             .as_ref()
             .map(|s| s.cpu_perc as f64)
             .unwrap_or(0.0),
@@ -361,12 +368,12 @@ pub async fn alert_servers(
               name: server.name.clone(),
               region: optional_string(&server.config.region),
               total_gb: server_status
-                .stats
+                .system_stats
                 .as_ref()
                 .map(|s| s.mem_total_gb)
                 .unwrap_or(0.0),
               used_gb: server_status
-                .stats
+                .system_stats
                 .as_ref()
                 .map(|s| s.mem_used_gb)
                 .unwrap_or(0.0),
@@ -388,12 +395,12 @@ pub async fn alert_servers(
             name: server.name.clone(),
             region: optional_string(&server.config.region),
             total_gb: server_status
-              .stats
+              .system_stats
               .as_ref()
               .map(|s| s.mem_total_gb)
               .unwrap_or(0.0),
             used_gb: server_status
-              .stats
+              .system_stats
               .as_ref()
               .map(|s| s.mem_used_gb)
               .unwrap_or(0.0),
@@ -409,12 +416,12 @@ pub async fn alert_servers(
           name: server.name.clone(),
           region: optional_string(&server.config.region),
           total_gb: server_status
-            .stats
+            .system_stats
             .as_ref()
             .map(|s| s.mem_total_gb)
             .unwrap_or(0.0),
           used_gb: server_status
-            .stats
+            .system_stats
             .as_ref()
             .map(|s| s.mem_used_gb)
             .unwrap_or(0.0),
@@ -452,7 +459,7 @@ pub async fn alert_servers(
             )
           {
             let disk =
-              server_status.stats.as_ref().and_then(|stats| {
+              server_status.system_stats.as_ref().and_then(|stats| {
                 stats.disks.iter().find(|disk| disk.mount == *path)
               });
             let alert = Alert {
@@ -487,7 +494,7 @@ pub async fn alert_servers(
           // modify alert level only if it has increased and not in maintenance
           if !in_maintenance && health.level < alert.level {
             let disk =
-              server_status.stats.as_ref().and_then(|stats| {
+              server_status.system_stats.as_ref().and_then(|stats| {
                 stats.disks.iter().find(|disk| disk.mount == *path)
               });
             alert.level = health.level;
@@ -505,9 +512,10 @@ pub async fn alert_servers(
         }
         (SeverityLevel::Ok, Some(alert), true) => {
           let mut alert = alert.clone();
-          let disk = server_status.stats.as_ref().and_then(|stats| {
-            stats.disks.iter().find(|disk| disk.mount == *path)
-          });
+          let disk =
+            server_status.system_stats.as_ref().and_then(|stats| {
+              stats.disks.iter().find(|disk| disk.mount == *path)
+            });
           alert.level = health.level;
           alert.data = AlertData::ServerDisk {
             id: server_status.id.clone(),

bin/core/src/monitor/alert/stack.rs

@@ -19,7 +19,7 @@ pub async fn alert_stacks(
 ) {
   let action_states = action_states();
   let mut alerts = Vec::<Alert>::new();
-  for status in stack_status_cache().get_list().await {
+  for status in stack_status_cache().get_values().await {
     // Don't alert if prev None
     let Some(prev) = status.prev else {
       continue;

bin/core/src/monitor/helpers.rs

@@ -7,11 +7,11 @@ use komodo_client::entities::{
   },
   repo::Repo,
   server::{
-    Server, ServerConfig, ServerHealth, ServerHealthState,
-    ServerState,
+    PeripheryInformation, Server, ServerConfig, ServerHealth,
+    ServerHealthState, ServerState,
   },
   stack::{ComposeProject, Stack, StackState},
-  stats::{SingleDiskUsage, SystemStats},
+  stats::{SingleDiskUsage, SystemInformation, SystemStats},
 };
 use serror::Serror;
 
@@ -103,20 +103,23 @@ type DockerLists = (
 pub async fn insert_server_status(
   server: &Server,
   state: ServerState,
-  version: String,
-  stats: Option<SystemStats>,
+  periphery_info: Option<PeripheryInformation>,
+  system_info: Option<SystemInformation>,
+  system_stats: Option<SystemStats>,
   (containers, networks, images, volumes, projects): DockerLists,
   err: impl Into<Option<Serror>>,
 ) {
-  let health = stats.as_ref().map(|s| get_server_health(server, s));
+  let health =
+    system_stats.as_ref().map(|s| get_server_health(server, s));
   server_status_cache()
     .insert(
       server.id.clone(),
       CachedServerStatus {
         id: server.id.clone(),
         state,
-        version,
-        stats,
+        periphery_info,
+        system_info,
+        system_stats,
         health,
         containers,
         networks,

bin/core/src/monitor/lists.rs

@@ -1,51 +0,0 @@
-use komodo_client::entities::{
-  docker::{
-    container::ContainerListItem, image::ImageListItem,
-    network::NetworkListItem, volume::VolumeListItem,
-  },
-  stack::ComposeProject,
-};
-use periphery_client::{
-  PeripheryClient,
-  api::{GetDockerLists, GetDockerListsResponse},
-};
-
-pub async fn get_docker_lists(
-  periphery: &PeripheryClient,
-) -> anyhow::Result<(
-  Vec<ContainerListItem>,
-  Vec<NetworkListItem>,
-  Vec<ImageListItem>,
-  Vec<VolumeListItem>,
-  Vec<ComposeProject>,
-)> {
-  let GetDockerListsResponse {
-    containers,
-    networks,
-    images,
-    volumes,
-    projects,
-  } = periphery.request(GetDockerLists {}).await?;
-  // TODO: handle the errors
-  let (
-    mut containers,
-    mut networks,
-    mut images,
-    mut volumes,
-    mut projects,
-  ) = (
-    containers.unwrap_or_default(),
-    networks.unwrap_or_default(),
-    images.unwrap_or_default(),
-    volumes.unwrap_or_default(),
-    projects.unwrap_or_default(),
-  );
-
-  containers.sort_by(|a, b| a.name.cmp(&b.name));
-  networks.sort_by(|a, b| a.name.cmp(&b.name));
-  images.sort_by(|a, b| a.name.cmp(&b.name));
-  volumes.sort_by(|a, b| a.name.cmp(&b.name));
-  projects.sort_by(|a, b| a.name.cmp(&b.name));
-
-  Ok((containers, networks, images, volumes, projects))
-}

bin/core/src/monitor/mod.rs

@@ -1,29 +1,37 @@
 use std::sync::{Arc, OnceLock};
 
 use async_timing_util::wait_until_timelength;
+use cache::CloneCache;
 use database::mungos::{find::find_collect, mongodb::bson::doc};
 use futures::future::join_all;
 use helpers::insert_stacks_status_unknown;
 use komodo_client::entities::{
-  deployment::DeploymentState,
+  build::Build,
+  deployment::{Deployment, DeploymentState},
   docker::{
     container::ContainerListItem, image::ImageListItem,
     network::NetworkListItem, volume::VolumeListItem,
   },
   komodo_timestamp, optional_string,
-  server::{Server, ServerHealth, ServerState},
-  stack::{ComposeProject, StackService, StackState},
-  stats::SystemStats,
+  repo::Repo,
+  server::{PeripheryInformation, Server, ServerHealth, ServerState},
+  stack::{ComposeProject, Stack, StackService, StackState},
+  stats::{SystemInformation, SystemStats},
+};
+use periphery_client::api::{
+  self, PollStatusResponse, git::GetLatestCommit,
 };
-use periphery_client::api::{self, git::GetLatestCommit};
 use serror::Serror;
 use tokio::sync::Mutex;
 
 use crate::{
   config::core_config,
-  helpers::{cache::Cache, periphery_client},
+  helpers::periphery_client,
   monitor::{alert::check_alerts, record::record_server_stats},
-  state::{db_client, deployment_status_cache, repo_status_cache},
+  state::{
+    db_client, deployment_status_cache, periphery_connections,
+    repo_status_cache,
+  },
 };
 
 use self::helpers::{
@@ -33,7 +41,6 @@ use self::helpers::{
 
 mod alert;
 mod helpers;
-mod lists;
 mod record;
 mod resources;
 
@@ -47,9 +54,10 @@ pub struct History<Curr: Default, Prev> {
 pub struct CachedServerStatus {
   pub id: String,
   pub state: ServerState,
-  pub version: String,
-  pub stats: Option<SystemStats>,
   pub health: Option<ServerHealth>,
+  pub periphery_info: Option<PeripheryInformation>,
+  pub system_info: Option<SystemInformation>,
+  pub system_stats: Option<SystemStats>,
   pub containers: Option<Vec<ContainerListItem>>,
   pub networks: Option<Vec<NetworkListItem>>,
   pub images: Option<Vec<ImageListItem>>,
@@ -122,8 +130,8 @@ async fn refresh_server_cache(ts: i64) {
 /// Makes sure cache for server doesn't update too frequently / simultaneously.
 /// If forced, will still block against simultaneous update.
 fn update_cache_for_server_controller()
--> &'static Cache<String, Arc<Mutex<i64>>> {
-  static CACHE: OnceLock<Cache<String, Arc<Mutex<i64>>>> =
+-> &'static CloneCache<String, Arc<Mutex<i64>>> {
+  static CACHE: OnceLock<CloneCache<String, Arc<Mutex<i64>>>> =
     OnceLock::new();
   CACHE.get_or_init(Default::default)
 }
@@ -154,64 +162,34 @@ pub async fn update_cache_for_server(server: &Server, force: bool) {
 
   *lock = now;
 
-  let (deployments, builds, repos, stacks) = tokio::join!(
-    find_collect(
-      &db_client().deployments,
-      doc! { "config.server_id": &server.id },
-      None,
-    ),
-    find_collect(&db_client().builds, doc! {}, None,),
-    find_collect(
-      &db_client().repos,
-      doc! { "config.server_id": &server.id },
-      None,
-    ),
-    find_collect(
-      &db_client().stacks,
-      doc! { "config.server_id": &server.id },
-      None,
-    )
-  );
-
-  let deployments =  deployments.inspect_err(|e| error!("failed to get deployments list from db (update status cache) | server : {} | {e:#}", server.name)).unwrap_or_default();
-  let builds =  builds.inspect_err(|e| error!("failed to get builds list from db (update status cache) | server : {} | {e:#}", server.name)).unwrap_or_default();
-  let repos = repos.inspect_err(|e|  error!("failed to get repos list from db (update status cache) | server: {} | {e:#}", server.name)).unwrap_or_default();
-  let stacks = stacks.inspect_err(|e|  error!("failed to get stacks list from db (update status cache) | server: {} | {e:#}", server.name)).unwrap_or_default();
+  let resources = UpdateCacheResources::load(server).await;
 
   // Handle server disabled
   if !server.config.enabled {
-    insert_deployments_status_unknown(deployments).await;
-    insert_stacks_status_unknown(stacks).await;
-    insert_repos_status_unknown(repos).await;
+    resources.insert_status_unknown().await;
     insert_server_status(
       server,
       ServerState::Disabled,
-      String::from("unknown"),
+      None,
+      None,
       None,
       (None, None, None, None, None),
       None,
     )
     .await;
+    periphery_connections().remove(&server.id).await;
     return;
   }
 
-  let Ok(periphery) = periphery_client(server) else {
-    error!(
-      "somehow periphery not ok to create. should not be reached."
-    );
-    return;
-  };
-
-  let version = match periphery.request(api::GetVersion {}).await {
-    Ok(version) => version.version,
+  let periphery = match periphery_client(server).await {
+    Ok(periphery) => periphery,
     Err(e) => {
-      insert_deployments_status_unknown(deployments).await;
-      insert_stacks_status_unknown(stacks).await;
-      insert_repos_status_unknown(repos).await;
+      resources.insert_status_unknown().await;
       insert_server_status(
         server,
         ServerState::NotOk,
-        String::from("Unknown"),
+        None,
+        None,
         None,
         (None, None, None, None, None),
         Serror::from(&e),
@@ -221,82 +199,75 @@ pub async fn update_cache_for_server(server: &Server, force: bool) {
     }
   };
 
-  let stats = if server.config.stats_monitoring {
-    match periphery.request(api::stats::GetSystemStats {}).await {
-      Ok(stats) => Some(filter_volumes(server, stats)),
-      Err(e) => {
-        insert_deployments_status_unknown(deployments).await;
-        insert_stacks_status_unknown(stacks).await;
-        insert_repos_status_unknown(repos).await;
-        insert_server_status(
-          server,
-          ServerState::NotOk,
-          String::from("unknown"),
-          None,
-          (None, None, None, None, None),
-          Serror::from(&e),
-        )
-        .await;
-        return;
-      }
+  let PollStatusResponse {
+    periphery_info,
+    system_info,
+    system_stats,
+    mut containers,
+    networks,
+    images,
+    volumes,
+    projects,
+  } = match periphery
+    .request(api::PollStatus {
+      include_stats: server.config.stats_monitoring,
+    })
+    .await
+  {
+    Ok(info) => info,
+    Err(e) => {
+      resources.insert_status_unknown().await;
+      insert_server_status(
+        server,
+        ServerState::NotOk,
+        None,
+        None,
+        None,
+        (None, None, None, None, None),
+        Serror::from(&e),
+      )
+      .await;
+      return;
     }
-  } else {
-    None
   };
 
-  match lists::get_docker_lists(&periphery).await {
-    Ok((mut containers, networks, images, volumes, projects)) => {
-      containers.iter_mut().for_each(|container| {
-        container.server_id = Some(server.id.clone())
-      });
-      tokio::join!(
-        resources::update_deployment_cache(
-          server.name.clone(),
-          deployments,
-          &containers,
-          &images,
-          &builds,
-        ),
-        resources::update_stack_cache(
-          server.name.clone(),
-          stacks,
-          &containers,
-          &images
-        ),
-      );
-      insert_server_status(
-        server,
-        ServerState::Ok,
-        version,
-        stats,
-        (
-          Some(containers.clone()),
-          Some(networks),
-          Some(images),
-          Some(volumes),
-          Some(projects),
-        ),
-        None,
-      )
-      .await;
-    }
-    Err(e) => {
-      insert_deployments_status_unknown(deployments).await;
-      insert_stacks_status_unknown(stacks).await;
-      insert_server_status(
-        server,
-        ServerState::Ok,
-        version,
-        stats,
-        (None, None, None, None, None),
-        Some(e.into()),
-      )
-      .await;
-    }
-  }
+  containers.iter_mut().for_each(|container| {
+    container.server_id = Some(server.id.clone())
+  });
+  tokio::join!(
+    resources::update_deployment_cache(
+      server.name.clone(),
+      resources.deployments,
+      &containers,
+      &images,
+      &resources.builds,
+    ),
+    resources::update_stack_cache(
+      server.name.clone(),
+      resources.stacks,
+      &containers,
+      &images
+    ),
+  );
+  insert_server_status(
+    server,
+    ServerState::Ok,
+    Some(periphery_info),
+    Some(system_info),
+    system_stats.map(|stats| filter_volumes(server, stats)),
+    (
+      Some(containers.clone()),
+      Some(networks),
+      Some(images),
+      Some(volumes),
+      Some(projects),
+    ),
+    None,
+  )
+  .await;
 
   let status_cache = repo_status_cache();
-  for repo in repos {
+  for repo in resources.repos {
     let (latest_hash, latest_message) = periphery
       .request(GetLatestCommit {
         name: repo.name.clone(),
@@ -320,6 +291,54 @@ pub async fn update_cache_for_server(server: &Server, force: bool) {
   }
 }
 
+struct UpdateCacheResources {
+  stacks: Vec<Stack>,
+  deployments: Vec<Deployment>,
+  builds: Vec<Build>,
+  repos: Vec<Repo>,
+}
+
+impl UpdateCacheResources {
+  pub async fn load(server: &Server) -> Self {
+    let (stacks, deployments, builds, repos) = tokio::join!(
+      find_collect(
+        &db_client().stacks,
+        doc! { "config.server_id": &server.id },
+        None,
+      ),
+      find_collect(
+        &db_client().deployments,
+        doc! { "config.server_id": &server.id },
+        None,
+      ),
+      find_collect(&db_client().builds, doc! {}, None,),
+      find_collect(
+        &db_client().repos,
+        doc! { "config.server_id": &server.id },
+        None,
+      ),
+    );
+
+    let stacks = stacks.inspect_err(|e|  error!("failed to get stacks list from db (update status cache) | server: {} | {e:#}", server.name)).unwrap_or_default();
+    let deployments =  deployments.inspect_err(|e| error!("failed to get deployments list from db (update status cache) | server : {} | {e:#}", server.name)).unwrap_or_default();
+    let builds =  builds.inspect_err(|e| error!("failed to get builds list from db (update status cache) | server : {} | {e:#}", server.name)).unwrap_or_default();
+    let repos = repos.inspect_err(|e|  error!("failed to get repos list from db (update status cache) | server: {} | {e:#}", server.name)).unwrap_or_default();
+
+    Self {
+      stacks,
+      deployments,
+      builds,
+      repos,
+    }
+  }
+
+  pub async fn insert_status_unknown(self) {
+    insert_stacks_status_unknown(self.stacks).await;
+    insert_deployments_status_unknown(self.deployments).await;
+    insert_repos_status_unknown(self.repos).await;
+  }
+}
+
 fn filter_volumes(
   server: &Server,
   mut stats: SystemStats,

bin/core/src/monitor/record.rs

@@ -6,11 +6,11 @@ use crate::state::{db_client, server_status_cache};
 
 #[instrument(level = "debug")]
 pub async fn record_server_stats(ts: i64) {
-  let status = server_status_cache().get_list().await;
+  let status = server_status_cache().get_values().await;
   let records = status
     .into_iter()
     .filter_map(|status| {
-      let stats = status.stats.as_ref()?;
+      let stats = status.system_stats.as_ref()?;
 
       let TotalDiskUsage {
         used_gb: disk_used_gb,

bin/core/src/periphery/mod.rs

@@ -0,0 +1,205 @@
+use std::{sync::Arc, time::Duration};
+
+use anyhow::{Context, anyhow};
+use bytes::Bytes;
+use cache::CloneCache;
+use periphery_client::api;
+use resolver_api::HasResponse;
+use serde::{Serialize, de::DeserializeOwned};
+use serde_json::json;
+use serror::deserialize_error_bytes;
+use tokio::sync::mpsc::{self, Sender};
+use tracing::warn;
+use transport::{
+  MessageState,
+  bytes::{from_transport_bytes, to_transport_bytes},
+};
+use uuid::Uuid;
+
+use crate::{
+  connection::{PeripheryConnection, PeripheryConnectionArgs},
+  state::periphery_connections,
+};
+
+pub mod terminal;
+
+pub type ConnectionChannels = CloneCache<Uuid, Sender<Bytes>>;
+
+#[derive(Debug)]
+pub struct PeripheryClient {
+  pub id: String,
+  channels: Arc<ConnectionChannels>,
+}
+
+impl PeripheryClient {
+  pub async fn new(
+    args: PeripheryConnectionArgs<'_>,
+    insecure_tls: bool,
+    // deprecated.
+    passkey: &str,
+  ) -> anyhow::Result<PeripheryClient> {
+    let connections = periphery_connections();
+
+    let id = args.id.to_string();
+
+    // Spawn client side connection if one doesn't exist.
+    let Some(connection) = connections.get(&id).await else {
+      if args.address.is_none() {
+        return Err(anyhow!("Server {id} is not connected"));
+      }
+      let channels = args
+        .spawn_client_connection(
+          id.clone(),
+          insecure_tls,
+          passkey.to_string(),
+        )
+        .await?;
+      return Ok(PeripheryClient { id, channels });
+    };
+
+    // Ensure the connection args are unchanged.
+    if args.matches(&connection.args) {
+      return Ok(PeripheryClient {
+        id,
+        channels: connection.channels.clone(),
+      });
+    }
+
+    // The args have changed.
+    if args.address.is_none() {
+      // Periphery -> Core connection
+      // Remove this connection, wait and see if client reconnects
+      connections.remove(&id).await;
+      tokio::time::sleep(Duration::from_millis(500)).await;
+      let connection = connections
+        .get(&id)
+        .await
+        .with_context(|| format!("Server {id} is not connected"))?;
+      Ok(PeripheryClient {
+        id,
+        channels: connection.channels.clone(),
+      })
+    } else {
+      // Core -> Periphery connection
+      let channels = args
+        .spawn_client_connection(
+          id.clone(),
+          insecure_tls,
+          passkey.to_string(),
+        )
+        .await?;
+      Ok(PeripheryClient { id, channels })
+    }
+  }
+
+  pub async fn cleanup(self) -> Option<Arc<PeripheryConnection>> {
+    periphery_connections().remove(&self.id).await
+  }
+
+  #[tracing::instrument(level = "debug", skip(self))]
+  pub async fn health_check(&self) -> anyhow::Result<()> {
+    self.request(api::GetHealth {}).await?;
+    Ok(())
+  }
+
+  #[tracing::instrument(
+    name = "PeripheryRequest",
+    skip(self),
+    level = "debug"
+  )]
+  pub async fn request<T>(
+    &self,
+    request: T,
+  ) -> anyhow::Result<T::Response>
+  where
+    T: std::fmt::Debug + Serialize + HasResponse,
+    T::Response: DeserializeOwned,
+  {
+    let connection =
+      periphery_connections().get(&self.id).await.with_context(
+        || format!("No connection found for server {}", self.id),
+      )?;
+
+    // Polls connected 3 times before bailing
+    connection.bail_if_not_connected().await?;
+
+    let id = Uuid::new_v4();
+    let (response_sender, mut response_receiever) =
+      mpsc::channel(1000);
+    self.channels.insert(id, response_sender).await;
+
+    let req_type = T::req_type();
+    let data = serde_json::to_vec(&json!({
+      "type": req_type,
+      "params": request
+    }))
+    .context("Failed to serialize request to bytes")?;
+
+    if let Err(e) = connection
+      .send(to_transport_bytes(data, id, MessageState::Request))
+      .await
+      .context("Failed to send request over channel")
+    {
+      // cleanup
+      self.channels.remove(&id).await;
+      return Err(e);
+    }
+
+    // Poll for the associated response
+    loop {
+      let next = tokio::select! {
+        msg = response_receiever.recv() => msg,
+        // Periphery will send InProgress every 5s to avoid timeout
+        _ = tokio::time::sleep(Duration::from_secs(10)) => {
+          return Err(anyhow!("Response timed out"));
+        }
+      };
+
+      let bytes = match next {
+        Some(bytes) => bytes,
+        None => {
+          return Err(anyhow!(
+            "Sender dropped before response was recieved"
+          ));
+        }
+      };
+
+      let (state, data) = match from_transport_bytes(bytes) {
+        Ok((data, _, state)) if !data.is_empty() => (state, data),
+        // Ignore no data cases
+        Ok(_) => continue,
+        Err(e) => {
+          warn!(
+            "Server {} | Received invalid message | {e:#}",
+            self.id
+          );
+          continue;
+        }
+      };
+      match state {
+        // TODO: improve the allocation in .to_vec
+        MessageState::Successful => {
+          // cleanup
+          self.channels.remove(&id).await;
+          return serde_json::from_slice(&data)
+            .context("Failed to parse successful response");
+        }
+        MessageState::Failed => {
+          // cleanup
+          self.channels.remove(&id).await;
+          return Err(deserialize_error_bytes(&data));
+        }
+        MessageState::InProgress => continue,
+        // Shouldn't be received by this receiver
+        other => {
+          // TODO: delete log
+          warn!(
+            "Server {} | Got other message over over response channel: {other:?}",
+            self.id
+          );
+          continue;
+        }
+      }
+    }
+  }
+}

bin/core/src/periphery/terminal.rs

@@ -0,0 +1,210 @@
+use std::{
+  pin::Pin,
+  sync::Arc,
+  task::{self, Poll},
+};
+
+use anyhow::Context;
+use bytes::Bytes;
+use cache::CloneCache;
+use futures::Stream;
+use periphery_client::api::terminal::{
+  ConnectContainerExec, ConnectTerminal, END_OF_OUTPUT,
+  ExecuteContainerExec, ExecuteTerminal,
+};
+use tokio::sync::mpsc::{Receiver, Sender, channel};
+use transport::bytes::data_from_transport_bytes;
+use uuid::Uuid;
+
+use crate::{
+  periphery::PeripheryClient, state::periphery_connections,
+};
+
+impl PeripheryClient {
+  pub async fn connect_terminal(
+    &self,
+    terminal: String,
+  ) -> anyhow::Result<(Uuid, Sender<Bytes>, Receiver<Bytes>)> {
+    tracing::trace!(
+      "request | type: ConnectTerminal | terminal name: {terminal}",
+    );
+
+    let connection =
+      periphery_connections().get(&self.id).await.with_context(
+        || format!("No connection found for server {}", self.id),
+      )?;
+
+    let id = self
+      .request(ConnectTerminal { terminal })
+      .await
+      .context("Failed to create terminal connection")?;
+
+    let (sender, receiever) = channel(1024);
+    connection.channels.insert(id, sender).await;
+
+    Ok((id, connection.sender.clone(), receiever))
+  }
+
+  pub async fn connect_container_exec(
+    &self,
+    container: String,
+    shell: String,
+  ) -> anyhow::Result<(Uuid, Sender<Bytes>, Receiver<Bytes>)> {
+    tracing::trace!(
+      "request | type: ConnectContainerExec | container name: {container} | shell: {shell}",
+    );
+
+    let connection =
+      periphery_connections().get(&self.id).await.with_context(
+        || format!("No connection found for server {}", self.id),
+      )?;
+
+    let id = self
+      .request(ConnectContainerExec { container, shell })
+      .await
+      .context("Failed to create container exec connection")?;
+
+    let (sender, receiever) = channel(1000);
+    connection.channels.insert(id, sender).await;
+
+    Ok((id, connection.sender.clone(), receiever))
+  }
+
+  /// Executes command on specified terminal,
+  /// and streams the response ending in [KOMODO_EXIT_CODE][komodo_client::entities::KOMODO_EXIT_CODE]
+  /// sentinal value as the expected final line of the stream.
+  ///
+  /// Example final line:
+  /// ```text
+  /// __KOMODO_EXIT_CODE:0
+  /// ```
+  ///
+  /// This means the command exited with code 0 (success).
+  ///
+  /// If this value is NOT the final item before stream closes, it means
+  /// the terminal exited mid command, before giving status. Example: running `exit`.
+  #[tracing::instrument(level = "debug", skip(self))]
+  pub async fn execute_terminal(
+    &self,
+    terminal: String,
+    command: String,
+  ) -> anyhow::Result<
+    impl Stream<Item = anyhow::Result<Bytes>> + 'static,
+  > {
+    tracing::trace!(
+      "sending request | type: ExecuteTerminal | terminal name: {terminal} | command: {command}",
+    );
+
+    let connection =
+      periphery_connections().get(&self.id).await.with_context(
+        || format!("No connection found for server {}", self.id),
+      )?;
+
+    let id = self
+      .request(ExecuteTerminal { terminal, command })
+      .await
+      .context("Failed to create execute terminal connection")?;
+
+    let (sender, receiver) = channel(1000);
+
+    connection.channels.insert(id, sender).await;
+
+    Ok(ReceiverStream {
+      id,
+      receiver,
+      channels: connection.channels.clone(),
+    })
+  }
+
+  /// Executes command on specified container,
+  /// and streams the response ending in [KOMODO_EXIT_CODE][komodo_client::entities::KOMODO_EXIT_CODE]
+  /// sentinal value as the expected final line of the stream.
+  ///
+  /// Example final line:
+  /// ```text
+  /// __KOMODO_EXIT_CODE:0
+  /// ```
+  ///
+  /// This means the command exited with code 0 (success).
+  ///
+  /// If this value is NOT the final item before stream closes, it means
+  /// the container shell exited mid command, before giving status. Example: running `exit`.
+  #[tracing::instrument(level = "debug", skip(self))]
+  pub async fn execute_container_exec(
+    &self,
+    container: String,
+    shell: String,
+    command: String,
+  ) -> anyhow::Result<ReceiverStream> {
+    tracing::trace!(
+      "sending request | type: ExecuteContainerExec | container: {container} | shell: {shell} | command: {command}",
+    );
+
+    let connection =
+      periphery_connections().get(&self.id).await.with_context(
+        || format!("No connection found for server {}", self.id),
+      )?;
+
+    let id = self
+      .request(ExecuteContainerExec {
+        container,
+        shell,
+        command,
+      })
+      .await
+      .context("Failed to create execute terminal connection")?;
+
+    let (sender, receiver) = channel(1000);
+
+    connection.channels.insert(id, sender).await;
+
+    Ok(ReceiverStream {
+      id,
+      receiver,
+      channels: connection.channels.clone(),
+    })
+  }
+}
+
+pub struct ReceiverStream {
+  id: Uuid,
+  channels: Arc<CloneCache<Uuid, Sender<Bytes>>>,
+  receiver: Receiver<Bytes>,
+}
+
+impl Stream for ReceiverStream {
+  type Item = anyhow::Result<Bytes>;
+  fn poll_next(
+    mut self: Pin<&mut Self>,
+    cx: &mut task::Context<'_>,
+  ) -> Poll<Option<Self::Item>> {
+    match self
+      .receiver
+      .poll_recv(cx)
+      .map(|bytes| bytes.map(data_from_transport_bytes))
+    {
+      Poll::Ready(Some(Ok(bytes))) if bytes == END_OF_OUTPUT => {
+        self.cleanup();
+        Poll::Ready(None)
+      }
+      Poll::Ready(Some(Ok(bytes))) => Poll::Ready(Some(Ok(bytes))),
+      Poll::Ready(Some(Err(e))) => Poll::Ready(Some(Err(e))),
+      Poll::Ready(None) => {
+        self.cleanup();
+        Poll::Ready(None)
+      }
+      Poll::Pending => Poll::Pending,
+    }
+  }
+}
+
+impl ReceiverStream {
+  fn cleanup(&self) {
+    // Not the prettiest but it should be fine
+    let channels = self.channels.clone();
+    let id = self.id;
+    tokio::spawn(async move {
+      channels.remove(&id).await;
+    });
+  }
+}

bin/core/src/resource/deployment.rs

@@ -267,7 +267,7 @@ impl super::KomodoResource for Deployment {
       );
       return Ok(());
     }
-    let periphery = match periphery_client(&server) {
+    let periphery = match periphery_client(&server).await {
       Ok(periphery) => periphery,
       Err(e) => {
         // This case won't ever happen, as periphery_client only fallible if the server is disabled.

bin/core/src/resource/mod.rs

@@ -79,6 +79,7 @@ pub use refresh::{
 pub use repo::{
   refresh_repo_state_cache, spawn_repo_state_refresh_loop,
 };
+pub use server::{rotate_server_keys, update_server_public_key};
 
 /// Implement on each Komodo resource for common methods
 pub trait KomodoResource {
@@ -240,10 +241,10 @@ pub async fn get<T: KomodoResource>(
   T::coll()
     .find_one(id_or_name_filter(id_or_name))
     .await
-    .context("failed to query db for resource")?
+    .context("Failed to query db for resource")?
     .with_context(|| {
       format!(
-        "did not find any {} matching {id_or_name}",
+        "Did not find any {} matching {id_or_name}",
         T::resource_type()
       )
     })
@@ -396,7 +397,7 @@ pub async fn list_full_for_user_using_document<T: KomodoResource>(
   )
   .await
   .with_context(|| {
-    format!("failed to pull {}s from mongo", T::resource_type())
+    format!("Failed to pull {}s from mongo", T::resource_type())
   })
 }
 
@@ -416,7 +417,7 @@ pub async fn get_id_to_resource_map<T: KomodoResource>(
   let res = find_collect(T::coll(), None, None)
     .await
     .with_context(|| {
-      format!("failed to pull {}s from mongo", T::resource_type())
+      format!("Failed to pull {}s from mongo", T::resource_type())
     })?
     .into_iter()
     .filter(|resource| {
@@ -459,6 +460,7 @@ pub async fn get_id_to_resource_map<T: KomodoResource>(
 pub async fn create<T: KomodoResource>(
   name: &str,
   mut config: T::PartialConfig,
+  info: Option<T::Info>,
   user: &User,
 ) -> serror::Result<Resource<T::Config, T::Info>> {
   if !T::user_can_create(user) {
@@ -517,7 +519,11 @@ pub async fn create<T: KomodoResource>(
     template: Default::default(),
     tags: Default::default(),
     config: config.into(),
-    info: T::default_info().await?,
+    info: if let Some(info) = info {
+      info
+    } else {
+      T::default_info().await?
+    },
     base_permission: PermissionLevel::None.into(),
     updated_at: start_ts,
   };
@@ -526,11 +532,11 @@ pub async fn create<T: KomodoResource>(
     .insert_one(&resource)
     .await
     .with_context(|| {
-      format!("failed to add {} to db", T::resource_type())
+      format!("Failed to add {} to db", T::resource_type())
     })?
     .inserted_id
     .as_object_id()
-    .context("inserted_id is not ObjectId")?
+    .context("Inserted_id is not ObjectId")?
     .to_string();
 
   let resource = get::<T>(&resource_id).await?;
@@ -547,18 +553,18 @@ pub async fn create<T: KomodoResource>(
   let mut update = make_update(target, T::create_operation(), user);
   update.start_ts = start_ts;
   update.push_simple_log(
-    &format!("create {}", T::resource_type()),
+    &format!("Create {}", T::resource_type()),
     format!(
-      "created {}\nid: {}\nname: {}",
+      "Created {}\nid: {}\nname: {}",
       T::resource_type(),
       resource.id,
       resource.name
     ),
   );
   update.push_simple_log(
-    "config",
+    "Config",
     serde_json::to_string_pretty(&resource.config)
-      .context("failed to serialize resource config to JSON")?,
+      .context("Failed to serialize resource config to JSON")?,
   );
 
   T::post_create(&resource, &mut update).await?;
@@ -754,7 +760,7 @@ pub async fn remove_tag_from_all<T: KomodoResource>(
   T::coll()
     .update_many(doc! {}, doc! { "$pull": { "tags": tag_id } })
     .await
-    .context("failed to remove tag from resources")?;
+    .context("Failed to remove tag from resources")?;
   Ok(())
 }
 
@@ -871,7 +877,7 @@ pub async fn delete<T: KomodoResource>(
     async {
       if let Err(e) = T::post_delete(&resource, &mut update).await {
         update
-          .push_error_log("post delete", format_serror(&e.into()));
+          .push_error_log("Post delete", format_serror(&e.into()));
       }
     },
     delete_from_alerters::<T>(&resource.id)
@@ -944,7 +950,7 @@ where
     .await
   {
     warn!(
-      "failed to delete_many permissions matching target {target:?} | {e:#}"
+      "Failed to delete_many permissions matching target {target:?} | {e:#}"
     );
   }
 }
@@ -979,7 +985,7 @@ where
       },
     )
     .await
-    .context("failed to remove resource from users recently viewed")
+    .context("Failed to remove resource from users recently viewed")
   {
     warn!("{e:#}");
   }

bin/core/src/resource/procedure.rs

@@ -774,6 +774,13 @@ async fn validate_config(
             ));
           }
         }
+        Execution::RotateAllServerKeys(_params) => {
+          if !user.admin {
+            return Err(anyhow!(
+              "Non admin user cannot trigger rotate all server keys"
+            ));
+          }
+        }
         Execution::Sleep(_) => {}
       }
     }

bin/core/src/resource/repo.rs

@@ -169,7 +169,7 @@ impl super::KomodoResource for Repo {
     }
 
     let server = super::get::<Server>(&repo.config.server_id).await?;
-    let periphery = periphery_client(&server)?;
+    let periphery = periphery_client(&server).await?;
 
     match periphery
       .request(DeleteRepo {

bin/core/src/resource/server.rs

@@ -1,30 +1,41 @@
+use std::str::FromStr;
+
 use anyhow::Context;
-use database::mungos::mongodb::{Collection, bson::doc};
+use database::mungos::mongodb::{
+  Collection,
+  bson::{doc, oid::ObjectId},
+};
 use indexmap::IndexSet;
 use komodo_client::entities::{
   Operation, ResourceTarget, ResourceTargetVariant, komodo_timestamp,
+  optional_string,
   permission::SpecificPermission,
   resource::Resource,
   server::{
     PartialServerConfig, Server, ServerConfig, ServerConfigDiff,
-    ServerListItem, ServerListItemInfo, ServerQuerySpecifics,
+    ServerInfo, ServerListItem, ServerListItemInfo,
+    ServerQuerySpecifics,
   },
   update::Update,
   user::User,
 };
+use periphery_client::api;
 
 use crate::{
   config::core_config,
-  helpers::query::get_system_info,
+  helpers::periphery_client,
   monitor::update_cache_for_server,
-  state::{action_states, db_client, server_status_cache},
+  state::{
+    action_states, db_client, periphery_connections,
+    server_status_cache,
+  },
 };
 
 impl super::KomodoResource for Server {
   type Config = ServerConfig;
   type PartialConfig = PartialServerConfig;
   type ConfigDiff = ServerConfigDiff;
-  type Info = ();
+  type Info = ServerInfo;
   type ListItem = ServerListItem;
   type QuerySpecifics = ServerQuerySpecifics;
 
@@ -57,11 +68,21 @@ impl super::KomodoResource for Server {
     server: Resource<Self::Config, Self::Info>,
   ) -> Self::ListItem {
     let status = server_status_cache().get(&server.id).await;
-    let (terminals_disabled, container_exec_disabled) =
-      get_system_info(&server)
-        .await
-        .map(|i| (i.terminals_disabled, i.container_exec_disabled))
-        .unwrap_or((true, true));
+    let (
+      version,
+      public_key,
+      terminals_disabled,
+      container_exec_disabled,
+    ) = match status.as_ref().and_then(|s| s.periphery_info.as_ref())
+    {
+      Some(info) => (
+        Some(info.version.clone()),
+        Some(info.public_key.clone()),
+        info.terminals_disabled,
+        info.container_exec_disabled,
+      ),
+      None => (None, None, true, true),
+    };
     ServerListItem {
       name: server.name,
       id: server.id,
@@ -70,12 +91,11 @@ impl super::KomodoResource for Server {
       resource_type: ResourceTargetVariant::Server,
       info: ServerListItemInfo {
         state: status.as_ref().map(|s| s.state).unwrap_or_default(),
-        version: status
-          .map(|s| s.version.clone())
-          .unwrap_or(String::from("Unknown")),
         region: server.config.region,
-        address: server.config.address,
-        external_address: server.config.external_address,
+        address: optional_string(server.config.address),
+        external_address: optional_string(
+          server.config.external_address,
+        ),
         send_unreachable_alerts: server
           .config
           .send_unreachable_alerts,
@@ -85,6 +105,11 @@ impl super::KomodoResource for Server {
         send_version_mismatch_alerts: server
           .config
           .send_version_mismatch_alerts,
+        version,
+        public_key,
+        attempted_public_key: optional_string(
+          server.info.attempted_public_key,
+        ),
         terminals_disabled,
         container_exec_disabled,
       },
@@ -212,6 +237,14 @@ impl super::KomodoResource for Server {
       .await
       .context("failed to close deleted server alerts")?;
 
+    let _ = db_client()
+      .onboarding_keys
+      .update_many(
+        doc! { "onboarded": &id },
+        doc! { "$pull": { "onboarded": &id } },
+      )
+      .await;
+
     Ok(())
   }
 
@@ -219,7 +252,41 @@ impl super::KomodoResource for Server {
     resource: &Resource<Self::Config, Self::Info>,
     _update: &mut Update,
   ) -> anyhow::Result<()> {
-    server_status_cache().remove(&resource.id).await;
+    tokio::join!(
+      server_status_cache().remove(&resource.id),
+      periphery_connections().remove(&resource.id),
+    );
     Ok(())
   }
 }
+
+pub async fn update_server_public_key(
+  server_id: &str,
+  public_key: &str,
+) -> anyhow::Result<()> {
+  db_client()
+    .servers
+    .update_one(
+      doc! { "_id": ObjectId::from_str(server_id)? },
+      doc! { "$set": { "info.public_key": public_key } },
+    )
+    .await
+    .context("Failed to update Server public key on database")?;
+  Ok(())
+}
+
+/// Rotates Periphery keys and updates
+/// `server.info.public_key` to match new public key.
+/// Does so without making a specific update.
+pub async fn rotate_server_keys(
+  server: &Server,
+) -> anyhow::Result<()> {
+  let periphery = periphery_client(&server).await?;
+  let public_key = periphery
+    .request(api::keys::RotatePrivateKey {})
+    .await
+    .context("Failed to rotate Periphery private key")?
+    .public_key;
+  update_server_public_key(&server.id, &public_key).await?;
+  Ok(())
+}

bin/core/src/resource/stack.rs

@@ -327,7 +327,7 @@ impl super::KomodoResource for Stack {
       return Ok(());
     }
 
-    let periphery = match periphery_client(&server) {
+    let periphery = match periphery_client(&server).await {
       Ok(periphery) => periphery,
       Err(e) => {
         // This case won't ever happen, as periphery_client only fallible if the server is disabled.

bin/core/src/stack/execute.rs

@@ -7,11 +7,12 @@ use komodo_client::{
     user::User,
   },
 };
-use periphery_client::{PeripheryClient, api::compose::*};
+use periphery_client::api::compose::*;
 
 use crate::{
   helpers::{periphery_client, update::update_update},
   monitor::update_cache_for_server,
+  periphery::PeripheryClient,
   state::action_states,
 };
 
@@ -55,7 +56,7 @@ pub async fn execute_compose<T: ExecuteCompose>(
   // Send update here for frontend to recheck action state
   update_update(update.clone()).await?;
 
-  let periphery = periphery_client(&server)?;
+  let periphery = periphery_client(&server).await?;
 
   if !services.is_empty() {
     update.logs.push(Log::simple(

bin/core/src/startup.rs

@@ -3,14 +3,17 @@ use std::str::FromStr;
 use colored::Colorize;
 use database::mungos::{
   find::find_collect,
-  mongodb::bson::{Document, doc, oid::ObjectId, to_document},
+  mongodb::bson::{
+    Document, doc, oid::ObjectId, to_bson, to_document,
+  },
 };
 use futures::future::join_all;
 use komodo_client::{
   api::{
     auth::SignUpLocalUser,
     execute::{
-      BackupCoreDatabase, Execution, GlobalAutoUpdate, RunAction,
+      BackupCoreDatabase, Execution, GlobalAutoUpdate,
+      RotateAllServerKeys, RunAction,
     },
     write::{
       CreateBuilder, CreateProcedure, CreateServer, CreateTag,
@@ -22,7 +25,7 @@ use komodo_client::{
     builder::{PartialBuilderConfig, PartialServerBuilderConfig},
     komodo_timestamp,
     procedure::{EnabledExecution, ProcedureConfig, ProcedureStage},
-    server::{PartialServerConfig, Server},
+    server::{PartialServerConfig, Server, ServerInfo},
     sync::ResourceSync,
     tag::TagColor,
     update::Log,
@@ -106,9 +109,10 @@ pub async fn on_startup() {
   tokio::join!(
     in_progress_update_cleanup(),
     open_alert_cleanup(),
-    clean_up_server_templates(),
     ensure_first_server_and_builder(),
     ensure_init_user_and_resources(),
+    clean_up_server_templates(),
+    init_server_info(),
   );
 }
 
@@ -198,51 +202,58 @@ async fn open_alert_cleanup() {
 /// Ensures a default server / builder exists with the defined address
 async fn ensure_first_server_and_builder() {
   let config = core_config();
-  let Some(address) = config.first_server.clone() else {
+  if config.first_server_name.is_none()
+    && config.first_server_address.is_none()
+  {
+    // If neither defined, early return
     return;
-  };
+  }
+  // Maybe create first Server / Builder
   let db = db_client();
-  let Ok(server) = db
-    .servers
-    .find_one(Document::new())
-    .await
-    .inspect_err(|e| error!("Failed to initialize 'first_server'. Failed to query db. {e:?}"))
+  // If any Server exists, exit early.
+  let Ok(None) =
+    db.servers.find_one(Document::new()).await.inspect_err(|e| {
+      error!(
+        "Failed to initialize first Server. Failed to query db. {e:?}"
+      )
+    })
   else {
     return;
   };
-  let server = if let Some(server) = server {
-    server
-  } else {
-    match (CreateServer {
-      name: config.first_server_name.clone(),
-      config: PartialServerConfig {
-        address: Some(address),
-        enabled: Some(true),
-        ..Default::default()
-      },
-    })
-    .resolve(&WriteArgs {
-      user: system_user().to_owned(),
-    })
-    .await
-    {
-      Ok(server) => server,
-      Err(e) => {
-        error!(
-          "Failed to initialize 'first_server'. Failed to CreateServer. {:#}",
-          e.error
-        );
-        return;
-      }
+  // Use the same name for Server and Builder
+  let name = config.first_server_name.as_deref().unwrap_or("Local");
+  let server = match (CreateServer {
+    name: name.to_string(),
+    config: PartialServerConfig {
+      address: config.first_server_address.clone(),
+      enabled: Some(true),
+      ..Default::default()
+    },
+    public_key: None,
+  })
+  .resolve(&WriteArgs {
+    user: system_user().to_owned(),
+  })
+  .await
+  {
+    Ok(server) => server,
+    Err(e) => {
+      error!(
+        "Failed to initialize first Server. Failed to CreateServer. {:#}",
+        e.error
+      );
+      return;
     }
   };
+  // If any builder exists, exit early.
   let Ok(None) = db.builders
     .find_one(Document::new()).await
     .inspect_err(|e| error!("Failed to initialize 'first_builder' | Failed to query db | {e:?}")) else {
       return;
     };
   if let Err(e) = (CreateBuilder {
-    name: config.first_server_name.clone(),
+    // Same name as Server
+    name: name.to_string(),
     config: PartialBuilderConfig::Server(
       PartialServerBuilderConfig {
         server_id: Some(server.id),
@@ -413,6 +424,58 @@ async fn ensure_init_user_and_resources() {
       );
     }
   }.await;
+
+  // RotateAllServerKeys
+  async {
+    let Ok(config) = ProcedureConfig::builder()
+      .stages(vec![ProcedureStage {
+        name: String::from("Stage 1"),
+        enabled: true,
+        executions: vec![
+          EnabledExecution {
+            execution: Execution::RotateAllServerKeys(RotateAllServerKeys {}),
+            enabled: true
+          }
+        ]
+      }])
+      .schedule(String::from("Every day at 06:00"))
+      .build()
+      .inspect_err(|e| error!("Failed to initialize Server key rotation Procedure | Failed to build Procedure | {e:?}")) else {
+      return;
+    };
+    let procedure = match (CreateProcedure {
+      name: String::from("Rotate Server Keys"),
+      config: config.into(),
+    })
+    .resolve(&write_args)
+    .await
+    {
+      Ok(procedure) => procedure,
+      Err(e) => {
+        error!(
+          "Failed to initialize Server key rotation Procedure | Failed to create Procedure | {:#}",
+          e.error
+        );
+        return;
+      }
+    };
+    if let Err(e) = (UpdateResourceMeta {
+      target: ResourceTarget::Procedure(procedure.id),
+      tags: Some(default_tags.clone()),
+      description: Some(String::from(
+        "Rotates all currently connected Server keys.",
+      )),
+      template: None,
+    })
+    .resolve(&write_args)
+    .await
+    {
+      warn!(
+        "Failed to update Server key rotation Procedure tags / description | {:#}",
+        e.error
+      );
+    }
+  }.await;
 }
 
 /// v1.17.5 removes the ServerTemplate resource.
@@ -457,3 +520,25 @@ async fn clean_up_server_templates() {
     },
   );
 }
+
+/// v2 adds ServerInfo to ServerSchema.
+/// Need to ensure it is initialized from null.
+async fn init_server_info() {
+  let default_server_info = match to_bson(&ServerInfo::default()) {
+    Ok(info) => info,
+    Err(e) => {
+      error!("Failed to serialize ServerInfo to bson | {e:?}");
+      return;
+    }
+  };
+  if let Err(e) = db_client()
+    .servers
+    .update_many(
+      doc! { "info": null },
+      doc! { "$set": { "info": default_server_info } },
+    )
+    .await
+  {
+    error!("Failed to migrate ServerInfo to v2 | {e:?}");
+  }
+}

bin/core/src/state.rs

@@ -5,7 +5,7 @@ use std::{
 
 use anyhow::Context;
 use arc_swap::ArcSwap;
-use database::Client;
+use cache::CloneCache;
 use komodo_client::entities::{
   action::ActionState,
   build::BuildState,
@@ -22,9 +22,9 @@ use octorust::auth::{
 use crate::{
   auth::jwt::JwtClient,
   config::core_config,
+  connection::PeripheryConnections,
   helpers::{
     action_state::ActionStates, all_resources::AllResourcesById,
-    cache::Cache,
   },
   monitor::{
     CachedDeploymentStatus, CachedRepoStatus, CachedServerStatus,
@@ -32,9 +32,9 @@ use crate::{
   },
 };
 
-static DB_CLIENT: OnceLock<Client> = OnceLock::new();
+static DB_CLIENT: OnceLock<database::Client> = OnceLock::new();
 
-pub fn db_client() -> &'static Client {
+pub fn db_client() -> &'static database::Client {
   DB_CLIENT
     .get()
     .expect("db_client accessed before initialized")
@@ -42,11 +42,13 @@ pub fn db_client() -> &'static Client {
 
 /// Must be called in app startup sequence.
 pub async fn init_db_client() {
-  let client = Client::new(&core_config().database)
+  let client = database::Client::new(&core_config().database)
     .await
     .context("failed to initialize database client")
     .unwrap();
-  DB_CLIENT.set(client).expect("db_clint");
+  DB_CLIENT
+    .set(client)
+    .expect("db_client initialized more than once");
 }
 
 pub fn jwt_client() -> &'static JwtClient {
@@ -60,6 +62,94 @@ pub fn jwt_client() -> &'static JwtClient {
   })
 }
 
+/// server id => connection
+pub fn periphery_connections() -> &'static PeripheryConnections {
+  static CONNECTIONS: OnceLock<PeripheryConnections> =
+    OnceLock::new();
+  CONNECTIONS.get_or_init(Default::default)
+}
+
+pub fn action_states() -> &'static ActionStates {
+  static ACTION_STATES: OnceLock<ActionStates> = OnceLock::new();
+  ACTION_STATES.get_or_init(ActionStates::default)
+}
+
+pub type ServerStatusCache =
+  CloneCache<String, Arc<CachedServerStatus>>;
+
+pub fn server_status_cache() -> &'static ServerStatusCache {
+  static SERVER_STATUS_CACHE: OnceLock<ServerStatusCache> =
+    OnceLock::new();
+  SERVER_STATUS_CACHE.get_or_init(Default::default)
+}
+
+pub type StackStatusCache =
+  CloneCache<String, Arc<History<CachedStackStatus, StackState>>>;
+
+pub fn stack_status_cache() -> &'static StackStatusCache {
+  static STACK_STATUS_CACHE: OnceLock<StackStatusCache> =
+    OnceLock::new();
+  STACK_STATUS_CACHE.get_or_init(Default::default)
+}
+
+/// Cache of ids to status
+pub type DeploymentStatusCache = CloneCache<
+  String,
+  Arc<History<CachedDeploymentStatus, DeploymentState>>,
+>;
+
+/// Cache of ids to status
+pub fn deployment_status_cache() -> &'static DeploymentStatusCache {
+  static DEPLOYMENT_STATUS_CACHE: OnceLock<DeploymentStatusCache> =
+    OnceLock::new();
+  DEPLOYMENT_STATUS_CACHE.get_or_init(Default::default)
+}
+
+pub type BuildStateCache = CloneCache<String, BuildState>;
+
+pub fn build_state_cache() -> &'static BuildStateCache {
+  static BUILD_STATE_CACHE: OnceLock<BuildStateCache> =
+    OnceLock::new();
+  BUILD_STATE_CACHE.get_or_init(Default::default)
+}
+
+pub type RepoStatusCache = CloneCache<String, Arc<CachedRepoStatus>>;
+
+pub fn repo_status_cache() -> &'static RepoStatusCache {
+  static REPO_STATUS_CACHE: OnceLock<RepoStatusCache> =
+    OnceLock::new();
+  REPO_STATUS_CACHE.get_or_init(Default::default)
+}
+
+pub type RepoStateCache = CloneCache<String, RepoState>;
+
+pub fn repo_state_cache() -> &'static RepoStateCache {
+  static REPO_STATE_CACHE: OnceLock<RepoStateCache> = OnceLock::new();
+  REPO_STATE_CACHE.get_or_init(Default::default)
+}
+
+pub type ProcedureStateCache = CloneCache<String, ProcedureState>;
+
+pub fn procedure_state_cache() -> &'static ProcedureStateCache {
+  static PROCEDURE_STATE_CACHE: OnceLock<ProcedureStateCache> =
+    OnceLock::new();
+  PROCEDURE_STATE_CACHE.get_or_init(Default::default)
+}
+
+pub type ActionStateCache = CloneCache<String, ActionState>;
+
+pub fn action_state_cache() -> &'static ActionStateCache {
+  static ACTION_STATE_CACHE: OnceLock<ActionStateCache> =
+    OnceLock::new();
+  ACTION_STATE_CACHE.get_or_init(Default::default)
+}
+
+pub fn all_resources_cache() -> &'static ArcSwap<AllResourcesById> {
+  static ALL_RESOURCES: OnceLock<ArcSwap<AllResourcesById>> =
+    OnceLock::new();
+  ALL_RESOURCES.get_or_init(Default::default)
+}
+
 pub fn github_client()
 -> Option<&'static HashMap<String, octorust::Client>> {
   static GITHUB_CLIENT: OnceLock<
@@ -129,83 +219,3 @@ pub fn github_client()
     })
     .as_ref()
 }
-
-pub fn action_states() -> &'static ActionStates {
-  static ACTION_STATES: OnceLock<ActionStates> = OnceLock::new();
-  ACTION_STATES.get_or_init(ActionStates::default)
-}
-
-/// Cache of ids to status
-pub type DeploymentStatusCache = Cache<
-  String,
-  Arc<History<CachedDeploymentStatus, DeploymentState>>,
->;
-
-/// Cache of ids to status
-pub fn deployment_status_cache() -> &'static DeploymentStatusCache {
-  static DEPLOYMENT_STATUS_CACHE: OnceLock<DeploymentStatusCache> =
-    OnceLock::new();
-  DEPLOYMENT_STATUS_CACHE.get_or_init(Default::default)
-}
-
-pub type StackStatusCache =
-  Cache<String, Arc<History<CachedStackStatus, StackState>>>;
-
-pub fn stack_status_cache() -> &'static StackStatusCache {
-  static STACK_STATUS_CACHE: OnceLock<StackStatusCache> =
-    OnceLock::new();
-  STACK_STATUS_CACHE.get_or_init(Default::default)
-}
-
-pub type ServerStatusCache = Cache<String, Arc<CachedServerStatus>>;
-
-pub fn server_status_cache() -> &'static ServerStatusCache {
-  static SERVER_STATUS_CACHE: OnceLock<ServerStatusCache> =
-    OnceLock::new();
-  SERVER_STATUS_CACHE.get_or_init(Default::default)
-}
-
-pub type RepoStatusCache = Cache<String, Arc<CachedRepoStatus>>;
-
-pub fn repo_status_cache() -> &'static RepoStatusCache {
-  static REPO_STATUS_CACHE: OnceLock<RepoStatusCache> =
-    OnceLock::new();
-  REPO_STATUS_CACHE.get_or_init(Default::default)
-}
-
-pub type BuildStateCache = Cache<String, BuildState>;
-
-pub fn build_state_cache() -> &'static BuildStateCache {
-  static BUILD_STATE_CACHE: OnceLock<BuildStateCache> =
-    OnceLock::new();
-  BUILD_STATE_CACHE.get_or_init(Default::default)
-}
-
-pub type RepoStateCache = Cache<String, RepoState>;
-
-pub fn repo_state_cache() -> &'static RepoStateCache {
-  static REPO_STATE_CACHE: OnceLock<RepoStateCache> = OnceLock::new();
-  REPO_STATE_CACHE.get_or_init(Default::default)
-}
-
-pub type ProcedureStateCache = Cache<String, ProcedureState>;
-
-pub fn procedure_state_cache() -> &'static ProcedureStateCache {
-  static PROCEDURE_STATE_CACHE: OnceLock<ProcedureStateCache> =
-    OnceLock::new();
-  PROCEDURE_STATE_CACHE.get_or_init(Default::default)
-}
-
-pub type ActionStateCache = Cache<String, ActionState>;
-
-pub fn action_state_cache() -> &'static ActionStateCache {
-  static ACTION_STATE_CACHE: OnceLock<ActionStateCache> =
-    OnceLock::new();
-  ACTION_STATE_CACHE.get_or_init(Default::default)
-}
-
-pub fn all_resources_cache() -> &'static ArcSwap<AllResourcesById> {
-  static ALL_RESOURCES: OnceLock<ArcSwap<AllResourcesById>> =
-    OnceLock::new();
-  ALL_RESOURCES.get_or_init(Default::default)
-}

bin/core/src/sync/execute.rs

@@ -144,6 +144,7 @@ pub trait ExecuteResourceSync: ResourceSyncTrait {
       let id = match crate::resource::create::<Self>(
         &resource.name,
         resource.config,
+        None,
         sync_user(),
       )
       .await

bin/core/src/sync/resources.rs

@@ -692,6 +692,7 @@ impl ResourceSyncTrait for Procedure {
           Execution::ClearRepoCache(_) => {}
           Execution::BackupCoreDatabase(_) => {}
           Execution::GlobalAutoUpdate(_) => {}
+          Execution::RotateAllServerKeys(_) => {}
           Execution::Sleep(_) => {}
         }
       }
@@ -822,6 +823,7 @@ impl ExecuteResourceSync for Procedure {
         let id = match crate::resource::create::<Procedure>(
           &name,
           resource.config.clone(),
+          None,
           sync_user(),
         )
         .await

bin/core/src/sync/toml.rs

@@ -810,7 +810,8 @@ impl ToToml for Procedure {
           | Execution::Sleep(_)
           | Execution::ClearRepoCache(_)
           | Execution::BackupCoreDatabase(_)
-          | Execution::GlobalAutoUpdate(_) => {}
+          | Execution::GlobalAutoUpdate(_)
+          | Execution::RotateAllServerKeys(_) => {}
         }
       }
     }

bin/core/src/ws/container.rs

@@ -21,7 +21,7 @@ pub async fn terminal(
 ) -> impl IntoResponse {
   ws.on_upgrade(|socket| async move {
     let Some((mut client_socket, user)) =
-      super::ws_login(socket).await
+      super::user_ws_login(socket).await
     else {
       return;
     };

bin/core/src/ws/deployment.rs

@@ -22,7 +22,7 @@ pub async fn terminal(
 ) -> impl IntoResponse {
   ws.on_upgrade(|socket| async move {
     let Some((mut client_socket, user)) =
-      super::ws_login(socket).await
+      super::user_ws_login(socket).await
     else {
       return;
     };

bin/core/src/ws/mod.rs

@@ -1,23 +1,29 @@
 use crate::{
   auth::{auth_api_key_check_enabled, auth_jwt_check_enabled},
   helpers::query::get_user,
+  periphery::PeripheryClient,
+  state::periphery_connections,
 };
 use anyhow::anyhow;
 use axum::{
   Router,
-  extract::ws::{CloseFrame, Message, Utf8Bytes, WebSocket},
+  extract::ws::{Message, WebSocket},
   routing::get,
 };
+use bytes::Bytes;
 use futures::{SinkExt, StreamExt};
 use komodo_client::{
   entities::{server::Server, user::User},
   ws::WsLoginMessage,
 };
-use tokio::net::TcpStream;
-use tokio_tungstenite::{
-  MaybeTlsStream, WebSocketStream, tungstenite,
-};
+use periphery_client::api::terminal::DisconnectTerminal;
+use tokio::sync::mpsc::{Receiver, Sender};
 use tokio_util::sync::CancellationToken;
+use transport::{
+  MessageState,
+  bytes::{data_from_transport_bytes, to_transport_bytes},
+};
+use uuid::Uuid;
 
 mod container;
 mod deployment;
@@ -27,6 +33,9 @@ mod update;
 
 pub fn router() -> Router {
   Router::new()
+    // Periphery facing
+    .route("/periphery", get(crate::connection::server::handler))
+    // User facing
     .route("/update", get(update::handler))
     .route("/terminal", get(terminal::handler))
     .route("/container/terminal", get(container::terminal))
@@ -35,7 +44,7 @@ pub fn router() -> Router {
 }
 
 #[instrument(level = "debug")]
-async fn ws_login(
+async fn user_ws_login(
   mut socket: WebSocket,
 ) -> Option<(WebSocket, User)> {
   let login_msg = match socket.recv().await {
@@ -131,7 +140,8 @@ async fn handle_container_terminal(
   container: String,
   shell: String,
 ) {
-  let periphery = match crate::helpers::periphery_client(server) {
+  let periphery = match crate::helpers::periphery_client(server).await
+  {
     Ok(periphery) => periphery,
     Err(e) => {
       debug!("couldn't get periphery | {e:#}");
@@ -145,34 +155,40 @@ async fn handle_container_terminal(
 
   trace!("connecting to periphery container exec websocket");
 
-  let periphery_socket = match periphery
-    .connect_container_exec(container, shell)
-    .await
-  {
-    Ok(ws) => ws,
-    Err(e) => {
-      debug!(
-        "Failed connect to periphery container exec websocket | {e:#}"
-      );
-      let _ = client_socket
-        .send(Message::text(format!("ERROR: {e:#}")))
-        .await;
-      let _ = client_socket.close().await;
-      return;
-    }
-  };
+  let (periphery_connection_id, periphery_sender, periphery_receiver) =
+    match periphery.connect_container_exec(container, shell).await {
+      Ok(ws) => ws,
+      Err(e) => {
+        debug!(
+          "Failed connect to periphery container exec websocket | {e:#}"
+        );
+        let _ = client_socket
+          .send(Message::text(format!("ERROR: {e:#}")))
+          .await;
+        let _ = client_socket.close().await;
+        return;
+      }
+    };
 
   trace!("connected to periphery container exec websocket");
 
-  core_periphery_forward_ws(client_socket, periphery_socket).await
+  forward_ws_channel(
+    periphery,
+    client_socket,
+    periphery_connection_id,
+    periphery_sender,
+    periphery_receiver,
+  )
+  .await
 }
 
-async fn core_periphery_forward_ws(
+async fn forward_ws_channel(
+  periphery: PeripheryClient,
   client_socket: axum::extract::ws::WebSocket,
-  periphery_socket: WebSocketStream<MaybeTlsStream<TcpStream>>,
+  periphery_connection_id: Uuid,
+  periphery_sender: Sender<Bytes>,
+  mut periphery_receiver: Receiver<Bytes>,
 ) {
-  let (mut periphery_send, mut periphery_receive) =
-    periphery_socket.split();
   let (mut core_send, mut core_receive) = client_socket.split();
   let cancel = CancellationToken::new();
 
@@ -188,15 +204,42 @@ async fn core_periphery_forward_ws(
         }
       };
       match res {
-        Some(Ok(msg)) => {
-          if let Err(e) =
-            periphery_send.send(axum_to_tungstenite(msg)).await
+        Some(Ok(Message::Binary(data))) => {
+          if let Err(e) = periphery_sender
+            .send(to_transport_bytes(
+              data.into(),
+              periphery_connection_id,
+              MessageState::Terminal,
+            ))
+            .await
           {
             debug!("Failed to send terminal message | {e:?}",);
             cancel.cancel();
             break;
           };
         }
+        Some(Ok(Message::Text(data))) => {
+          let data: Bytes = data.into();
+          if let Err(e) = periphery_sender
+            .send(to_transport_bytes(
+              data.into(),
+              periphery_connection_id,
+              MessageState::Terminal,
+            ))
+            .await
+          {
+            debug!("Failed to send terminal message | {e:?}",);
+            cancel.cancel();
+            break;
+          };
+        }
+        // TODO: Disconnect from periphery when client disconnects
+        Some(Ok(Message::Close(_frame))) => {
+          cancel.cancel();
+          break;
+        }
+        // Ignore
+        Some(Ok(_)) => {}
         Some(Err(_e)) => {
           cancel.cancel();
           break;
@@ -212,31 +255,23 @@ async fn core_periphery_forward_ws(
   let periphery_to_core = async {
     loop {
       let res = tokio::select! {
-        res = periphery_receive.next() => res,
+        res = periphery_receiver.recv() => res.map(data_from_transport_bytes),
         _ = cancel.cancelled() => {
           trace!("periphery to core read: cancelled from inside");
           break;
         }
       };
       match res {
-        Some(Ok(msg)) => {
-          if let Err(e) =
-            core_send.send(tungstenite_to_axum(msg)).await
+        Some(Ok(bytes)) => {
+          if let Err(e) = core_send.send(Message::Binary(bytes)).await
           {
             debug!("{e:?}");
             cancel.cancel();
             break;
           };
         }
-        Some(Err(e)) => {
-          let _ = core_send
-              .send(Message::text(format!(
-                "ERROR: Failed to receive message from periphery | {e:?}"
-              )))
-              .await;
-          cancel.cancel();
-          break;
-        }
+        // No data, ignore
+        Some(Err(_e)) => {}
         None => {
           let _ = core_send.send(Message::text("STREAM EOF")).await;
           cancel.cancel();
@@ -247,44 +282,21 @@ async fn core_periphery_forward_ws(
   };
 
   tokio::join!(core_to_periphery, periphery_to_core);
-}
 
-fn axum_to_tungstenite(msg: Message) -> tungstenite::Message {
-  match msg {
-    Message::Text(text) => tungstenite::Message::Text(
-      // TODO: improve this conversion cost from axum ws library
-      tungstenite::Utf8Bytes::from(text.to_string()),
-    ),
-    Message::Binary(bytes) => tungstenite::Message::Binary(bytes),
-    Message::Ping(bytes) => tungstenite::Message::Ping(bytes),
-    Message::Pong(bytes) => tungstenite::Message::Pong(bytes),
-    Message::Close(close_frame) => {
-      tungstenite::Message::Close(close_frame.map(|cf| {
-        tungstenite::protocol::CloseFrame {
-          code: cf.code.into(),
-          reason: tungstenite::Utf8Bytes::from(cf.reason.to_string()),
-        }
-      }))
-    }
-  }
-}
-
-fn tungstenite_to_axum(msg: tungstenite::Message) -> Message {
-  match msg {
-    tungstenite::Message::Text(text) => {
-      Message::Text(Utf8Bytes::from(text.to_string()))
-    }
-    tungstenite::Message::Binary(bytes) => Message::Binary(bytes),
-    tungstenite::Message::Ping(bytes) => Message::Ping(bytes),
-    tungstenite::Message::Pong(bytes) => Message::Pong(bytes),
-    tungstenite::Message::Close(close_frame) => {
-      Message::Close(close_frame.map(|cf| CloseFrame {
-        code: cf.code.into(),
-        reason: Utf8Bytes::from(cf.reason.to_string()),
-      }))
-    }
-    tungstenite::Message::Frame(_) => {
-      unreachable!()
-    }
+  // Cleanup
+  if let Err(e) = periphery
+    .request(DisconnectTerminal {
+      id: periphery_connection_id,
+    })
+    .await
+  {
+    warn!(
+      "Failed to disconnect Periphery terminal forwarding | {e:#}",
+    )
+  }
+  if let Some(connection) =
+    periphery_connections().get(&periphery.id).await
+  {
+    connection.channels.remove(&periphery_connection_id).await;
   }
 }

bin/core/src/ws/stack.rs

@@ -26,7 +26,7 @@ pub async fn terminal(
 ) -> impl IntoResponse {
   ws.on_upgrade(|socket| async move {
     let Some((mut client_socket, user)) =
-      super::ws_login(socket).await
+      super::user_ws_login(socket).await
     else {
       return;
     };

bin/core/src/ws/terminal.rs

@@ -10,7 +10,7 @@ use komodo_client::{
 
 use crate::{
   helpers::periphery_client, permission::get_check_permissions,
-  ws::core_periphery_forward_ws,
+  ws::forward_ws_channel,
 };
 
 #[instrument(name = "ConnectTerminal", skip(ws))]
@@ -22,7 +22,7 @@ pub async fn handler(
 ) -> impl IntoResponse {
   ws.on_upgrade(|socket| async move {
     let Some((mut client_socket, user)) =
-      super::ws_login(socket).await
+      super::user_ws_login(socket).await
     else {
       return;
     };
@@ -45,7 +45,7 @@ pub async fn handler(
       }
     };
 
-    let periphery = match periphery_client(&server) {
+    let periphery = match periphery_client(&server).await {
       Ok(periphery) => periphery,
       Err(e) => {
         debug!("couldn't get periphery | {e:#}");
@@ -59,21 +59,31 @@ pub async fn handler(
 
     trace!("connecting to periphery terminal websocket");
 
-    let periphery_socket =
-      match periphery.connect_terminal(terminal).await {
-        Ok(ws) => ws,
-        Err(e) => {
-          debug!("Failed connect to periphery terminal | {e:#}");
-          let _ = client_socket
-            .send(Message::text(format!("ERROR: {e:#}")))
-            .await;
-          let _ = client_socket.close().await;
-          return;
-        }
-      };
+    let (
+      periphery_connection_id,
+      periphery_sender,
+      periphery_receiver,
+    ) = match periphery.connect_terminal(terminal).await {
+      Ok(ws) => ws,
+      Err(e) => {
+        debug!("Failed connect to periphery terminal | {e:#}");
+        let _ = client_socket
+          .send(Message::text(format!("ERROR: {e:#}")))
+          .await;
+        let _ = client_socket.close().await;
+        return;
+      }
+    };
 
     trace!("connected to periphery terminal websocket");
 
-    core_periphery_forward_ws(client_socket, periphery_socket).await
+    forward_ws_channel(
+      periphery,
+      client_socket,
+      periphery_connection_id,
+      periphery_sender,
+      periphery_receiver,
+    )
+    .await
   })
 }

bin/core/src/ws/update.rs

@@ -23,7 +23,7 @@ pub async fn handler(ws: WebSocketUpgrade) -> impl IntoResponse {
 
   // handle http -> ws updgrade
   ws.on_upgrade(|socket| async move {
-    let Some((socket, user)) = super::ws_login(socket).await else {
+    let Some((socket, user)) = super::user_ws_login(socket).await else {
       return
     };
 

bin/periphery/Cargo.toml

@@ -21,11 +21,13 @@ environment_file.workspace = true
 environment.workspace = true
 interpolate.workspace = true
 formatting.workspace = true
+transport.workspace = true
 response.workspace = true
 command.workspace = true
 config.workspace = true
 logger.workspace = true
 cache.workspace = true
+noise.workspace = true
 git.workspace = true
 # mogh
 serror = { workspace = true, features = ["axum"] }
@@ -34,12 +36,13 @@ derive_variants.workspace = true
 resolver_api.workspace = true
 run_command.workspace = true
 # external
-pin-project-lite.workspace = true
+serde_yaml_ng.workspace = true
 tokio-stream.workspace = true
 portable-pty.workspace = true
+shell-escape.workspace = true
 axum-server.workspace = true
+urlencoding.workspace = true
 serde_json.workspace = true
-serde_yaml_ng.workspace = true
 tokio-util.workspace = true
 arc-swap.workspace = true
 colored.workspace = true
@@ -57,5 +60,3 @@ axum.workspace = true
 clap.workspace = true
 envy.workspace = true
 uuid.workspace = true
-rand.workspace = true
-shell-escape.workspace = true

bin/periphery/aio.Dockerfile

@@ -1,6 +1,6 @@
 ## All in one, multi stage compile + runtime Docker build for your architecture.
 
-FROM rust:1.89.0-bullseye AS builder
+FROM rust:1.90.0-bullseye AS builder
 RUN cargo install cargo-strip
 
 WORKDIR /builder
@@ -24,8 +24,13 @@ COPY --from=builder /builder/target/release/periphery /usr/local/bin/periphery
 
 EXPOSE 8120
 
+# Can mount config file to /config/*config*.toml and it will be picked up.
+ENV PERIPHERY_CONFIG_PATHS="/config"
+# Change the default in container to /config/keys to match Core
+ENV PERIPHERY_PRIVATE_KEY="file:/config/keys/periphery.key"
+
 CMD [ "periphery" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Periphery"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/periphery/multi-arch.Dockerfile

@@ -27,8 +27,13 @@ RUN mv /app/arch/${TARGETPLATFORM} /usr/local/bin/periphery && rm -r /app/arch
 
 EXPOSE 8120
 
+# Can mount config file to /config/*config*.toml and it will be picked up.
+ENV PERIPHERY_CONFIG_PATHS="/config"
+# Change the default in container to /config/keys to match Core
+ENV PERIPHERY_PRIVATE_KEY="file:/config/keys/periphery.key"
+
 CMD [ "periphery" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Periphery"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/periphery/single-arch.Dockerfile

@@ -16,8 +16,13 @@ COPY --from=binaries /periphery /usr/local/bin/periphery
 
 EXPOSE 8120
 
+# Can mount config file to /config/*config*.toml and it will be picked up.
+ENV PERIPHERY_CONFIG_PATHS="/config"
+# Change the default in container to /config/keys to match Core
+ENV PERIPHERY_PRIVATE_KEY="file:/config/keys/periphery.key"
+
 CMD [ "periphery" ]
 
-LABEL org.opencontainers.image.source=https://github.com/moghtech/komodo
+LABEL org.opencontainers.image.source="https://github.com/moghtech/komodo"
 LABEL org.opencontainers.image.description="Komodo Periphery"
-LABEL org.opencontainers.image.licenses=GPL-3.0
+LABEL org.opencontainers.image.licenses="GPL-3.0"

bin/periphery/src/api/build/mod.rs

@@ -25,12 +25,14 @@ use resolver_api::Resolve;
 use tokio::fs;
 
 use crate::{
-  build::{parse_build_args, parse_secret_args, write_dockerfile},
   config::periphery_config,
   docker::docker_login,
-  helpers::{parse_extra_args, parse_labels},
+  helpers::{format_extra_args, format_labels},
 };
 
+mod helpers;
+use helpers::*;
+
 impl Resolve<super::Args> for GetDockerfileContentsOnHost {
   #[instrument(name = "GetDockerfileContentsOnHost", level = "debug")]
   async fn resolve(
@@ -278,11 +280,11 @@ impl Resolve<super::Args> for build::Build {
     let command_secret_args =
       parse_secret_args(&secret_args, &build_path).await?;
 
-    let labels = parse_labels(
+    let labels = format_labels(
       &environment_vars_from_str(labels).context("Invalid labels")?,
     );
 
-    let extra_args = parse_extra_args(extra_args);
+    let extra_args = format_extra_args(extra_args);
 
     let buildx = if *use_buildx { " buildx" } else { "" };
 

bin/periphery/src/api/compose/helpers.rs

@@ -1,26 +1,55 @@
-use std::{fmt::Write, path::PathBuf};
+use std::{
+  fmt::Write,
+  path::{Path, PathBuf},
+};
 
 use anyhow::{Context, anyhow};
 use command::run_komodo_command;
+use formatting::format_serror;
 use komodo_client::entities::{
-  RepoExecutionArgs, repo::Repo, stack::Stack,
+  FileContents, RepoExecutionArgs,
+  repo::Repo,
+  stack::{Stack, StackRemoteFileContents},
   to_path_compatible_name,
+  update::Log,
 };
 use periphery_client::api::{
   compose::ComposeUpResponse, git::PullOrCloneRepo,
 };
 use resolver_api::Resolve;
+use tokio::fs;
 
-use crate::config::periphery_config;
+use crate::{
+  api::Args, config::periphery_config, docker::docker_login,
+};
 
-pub mod up;
-pub mod write;
+use super::docker_compose;
 
-pub fn docker_compose() -> &'static str {
-  if periphery_config().legacy_compose_cli {
-    "docker-compose"
-  } else {
-    "docker compose"
+pub async fn maybe_login_registry(
+  stack: &Stack,
+  registry_token: Option<String>,
+  logs: &mut Vec<Log>,
+) {
+  if !stack.config.registry_provider.is_empty()
+    && !stack.config.registry_account.is_empty()
+    && let Err(e) = docker_login(
+      &stack.config.registry_provider,
+      &stack.config.registry_account,
+      registry_token.as_deref(),
+    )
+    .await
+    .with_context(|| {
+      format!(
+        "Domain: '{}' | Account: '{}'",
+        stack.config.registry_provider, stack.config.registry_account
+      )
+    })
+    .context("Failed to login to image registry")
+  {
+    logs.push(Log::error(
+      "Login to Registry",
+      format_serror(&e.into()),
+    ));
   }
 }
 
@@ -53,7 +82,7 @@ pub fn env_file_args(
   Ok(res)
 }
 
-pub async fn down(
+pub async fn compose_down(
   project: &str,
   services: &[String],
   res: &mut ComposeUpResponse,
@@ -89,6 +118,7 @@ pub async fn pull_or_clone_stack(
   stack: &Stack,
   repo: Option<&Repo>,
   git_token: Option<String>,
+  req_args: &Args,
 ) -> anyhow::Result<PathBuf> {
   if stack.config.files_on_host {
     return Err(anyhow!(
@@ -135,9 +165,78 @@ pub async fn pull_or_clone_stack(
     skip_secret_interp: Default::default(),
     replacers: Default::default(),
   }
-  .resolve(&crate::api::Args)
+  .resolve(req_args)
   .await
   .map_err(|e| e.error)?;
 
   Ok(root)
 }
+
+pub async fn validate_files(
+  stack: &Stack,
+  run_directory: &Path,
+  res: &mut ComposeUpResponse,
+) {
+  let file_paths = stack
+    .all_file_dependencies()
+    .into_iter()
+    .map(|file| {
+      (
+        // This will remove any intermediate uneeded '/./' in the path
+        run_directory
+          .join(&file.path)
+          .components()
+          .collect::<PathBuf>(),
+        file,
+      )
+    })
+    .collect::<Vec<_>>();
+
+  // First validate no missing files
+  for (full_path, file) in &file_paths {
+    if !full_path.exists() {
+      res.missing_files.push(file.path.clone());
+    }
+  }
+  if !res.missing_files.is_empty() {
+    res.logs.push(Log::error(
+      "Validate Files",
+      format_serror(
+        &anyhow!(
+          "Missing files: {}", res.missing_files.join(", ")
+        )
+        .context("Ensure the run_directory and all file paths are correct.")
+        .context("A file doesn't exist after writing stack.")
+        .into(),
+      ),
+    ));
+    return;
+  }
+
+  for (full_path, file) in file_paths {
+    let file_contents =
+      match fs::read_to_string(&full_path).await.with_context(|| {
+        format!("Failed to read file contents at {full_path:?}")
+      }) {
+        Ok(res) => res,
+        Err(e) => {
+          let error = format_serror(&e.into());
+          res
+            .logs
+            .push(Log::error("Read Compose File", error.clone()));
+          // This should only happen for repo stacks, ie remote error
+          res.remote_errors.push(FileContents {
+            path: file.path,
+            contents: error,
+          });
+          return;
+        }
+      };
+    res.file_contents.push(StackRemoteFileContents {
+      path: file.path,
+      contents: file_contents,
+      services: file.services,
+      requires: file.requires,
+    });
+  }
+}