komodo/changelog.md

Changelog

Monitor v1.13 - Komodo (Aug 2024)

• This is the first named release, as I think it is really big. The Komodo Dragon is the largest species of Monitor lizard.
• Deploy docker compose with the new Stack resource.
  • Can define the compose file in the UI, or direct Monitor to clone a git repo containing compose files.
  • Use webhooks to redeploy the stack on push to the repo
  • Manage the environment variables passed to the compose command.
• Builds can now be configured with an alternate repository name to push the image under. -An optional tag can also be configured to be postfixed onto the version, like image:1.13-aarch64. This helps for pushing alternate build configurations under the same image repo, just under different tags.
• Repos can now be "built" using builders. The idea is, you spawn an AWS instance, clone a repo, execute a shell command (like running a script in the repo), and terminating the instance. The script can build a binary, and push it to some binary repository. Users will have to manage their own versioning though.
• High level UI Updates courtesy of @karamvirsingh98

v1.12 (July 2024)

• Break free of Github dependance. Use other git providers, including self hosted ones.
• Same for Docker registry. You can also now use any docker registry for your images.

v1 (Spring 2024)

• New resource types:

  • Repo: Clone / pull configured repositories on desired Server. Run shell commands in the repo on every clone / pull to acheive automation. Listen for pushes to a particular branch to automatically pull the repo and run the command.
  • Procedure: Combine multiple executions (Deploy, RunBuild) and run them in sequence or in parallel. RunProcedure is an execution type, meaning procedures can run other procedures.
  • Builder: Ephemeral builder configuration has moved to being an API / UI managed entity for greater observability and ease of management.
  • Alerter: Define multiple alerting endpoints and manage them via the API / UI.
    • Slack support continues with the Slack Alerter variant.
    • Send JSON serialized alert data to any HTTP endpoint with the Custom Alerter variant.
  • Template: Define a template for your cloud provider's VM configuration
    • Launch VMs based on the template and automatically add them as Monitor servers.
    • Supports AWS EC2 and Hetzner Cloud.
  • Sync: Sync resources declared in toml files in Github repos.
    • Manage resources declaratively, with git history for configuration rollbacks.
    • See the actions which will be performed in the UI, and execute them upon manual confirmation.
    • Use a Github webhook to automatically execute syncs on git push.

• Resource Tagging

  • Attach multiple tags to resources, which can be used to group related resources together. These can be used to filter resources in the UI.
  • For example, resources can be given tags for environment, like Prod, Uat, or Dev. This can be combined with tags for the larger system the resource is a part of, such as Authentication, Logging, or Messaging.
  • Proper tagging will make it easy to find resources and visualize system components, even as the number of resources grows large.

• Variables

  • Manage global, non-secret key-value pairs using the API / UI.
  • These values can be interpolated into deployment environments and build build_args

• Core Accounts and Secrets

  • Docker / Github accounts and Secrets can now be configured in the Core configuration file.
  • They can still be added to the Periphery configuration as before. Accounts / Secrets defined in the Core configuration will be preferentially used over like ones defined in Periphery configuration.

• User Groups

  • Admins can now create User Groups and assign permissions to them as if they were a user.
  • Multiple users can then be added to the group, and a user can be added to multiple groups.
  • Users in the group inherit the group's permissions.

• Builds

  • Build log displays the latest commit hash and message.
  • In-progress builds are able to be cancelled before completion.
  • Specify a specific commit hash to build from.

• Deployments

  • Filter log lines using multiple search terms.

• Alerting

  • The alerting system has been redesigned with a stateful model.
  • Alerts can be in an Open or a Resolved state, and alerts are only sent on state changes.
  • For example, say a server has just crossed 80% memory usage, the configured memory threshold. An alert will be created in the Open state and the alert data will be sent out. Later, it has dropped down to 70%. The alert will be changed to the Resolved state and the alert data will again be sent.
  • In addition to server usage alerts, Monitor now supports deployment state change alerts. These are sent when a deployment's state changes without being caused by a Monitor action. For example, if a deployment goes from the Running state to the Exited state unexpectedly, say from a crash, an alert will be sent.
  • Current and past alerts can be retrieved using the API and viewed on the UI.

• New UI:

  • The Monitor UI has been revamped to support the new features and improve the user experience.

v0 (Winter 2022)

• Move Core and Periphery implementation to Rust.
• Add AWS Ec2 ephemeral build instance support.
  • Configuration added to core configuration file.
• Automatic build versioning system, supporting image rollback.
• Realtime and historical system stats - CPU, memory, disk.
• Simple stats alerting based on out-of-bounds values for system stats.
  • Support sending alerts to Slack.

Pre-versioned releases

• Defined main resource types:
  • Server
  • Deployment
  • Build
• Basics of Monitor:
  • Build docker images from Github repos.
  • Manage image deployment on connected servers, see container status, get container logs.
  • Add account credentials in Periphery configuration.
• Core and Periphery implemented in Typescript.