mirror of
https://github.com/fosrl/pangolin.git
synced 2026-05-08 13:49:15 -05:00
[GH-ISSUE #2506] Overnight, all requests to pangolin port 443 return 404 #2176
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @bernhardkaindl on GitHub (Feb 19, 2026).
Original GitHub issue: https://github.com/fosrl/pangolin/issues/2506
Describe the Bug
Logs look similar to #2346
[Traefik unable to reach "http://pangolin:3001/api/v1/traefik-config\"], cc @SteelyxYT(closed by the bot automatically due to inactivity)
Stock pangolin on a VPS of Oracle Frankfurt(Germany), installed using the install script on January 27 with crowdsec
It looks like it starts when trying to do ACME renew with
It looks like the VPS had lost connectivity to traefik servers, restarting the
docker compose restartrecovered the pangolin stack:Environment
To Reproduce
Appeared overnight without any action.
Expected Behavior
Continue serving requests as is.
@SteelyxYT commented on GitHub (Feb 19, 2026):
Does your site still return the dashboard? Personally I got 404 on all except the pangolin dashboard.
@xupefei commented on GitHub (Feb 19, 2026):
I got the same issue after my cloud provider restarted my VPS. All resources are inaccessible (Chrome message: CONNECTION_REFUSED), including the Pangolin dashboard. I have to run
docker compose restartto bring it back online.Here's the error message from
gerbil:@miloschwartz commented on GitHub (Feb 25, 2026):
Are containers crashing or locking up? This stuff tends to happen when on extremely resource constrained VPS and things start to lock up. The errors indicate the Pangolin and/or Gerbil containers become inaccessible while Traefik stays online. Manually check if they go unresponsive next time and see if there are any logs from these containers to indicate something is up
@bernhardkaindl commented on GitHub (Mar 2, 2026):
In my case, the containers themselves showed no errors, and the server had plenty of spare memory. The containers consume 3GB of memory, and the VPS had at least 12GB, I think even 24GB of RAM at the time. I checked some of the containers to see if they were responding using
curl.@esoadamo commented on GitHub (Mar 2, 2026):
I had similar issue a few days back when everything started returning 404 for no apparent reason. I am running Pangolin's Traefik behind a Caddy reverse proxy and had to enable SNI to be sent to the Traefik as well. This is my Caddy config now:
@github-actions[bot] commented on GitHub (Mar 17, 2026):
This issue has been automatically marked as stale due to 14 days of inactivity. It will be closed in 14 days if no further activity occurs.
@github-actions[bot] commented on GitHub (Mar 31, 2026):
This issue has been automatically closed due to inactivity. If you believe this is still relevant, please open a new issue with up-to-date information.
@m-elsharkawi commented on GitHub (Apr 13, 2026):
I found what was causing a similar issue for me. Docker bridge networking on the host was broken by stale nftables raw PREROUTING rules left behind after the outage and network recreation.
Symptoms were:
In my case,
pangolinandgerbilwere on the same bridge, butnft list rulesetshowed old rules like these still dropping traffic:Those rules were stale and no longer matched the current bridge.
What fixed it:
Identify the stale drop rules for the affected container IPs.
Delete the bad rules by handle:
In my case, removing the stale rules immediately restored container-to-container connectivity and Pangolin started working again.
A Docker restart alone did not remove those stale rules, so checking
nftdirectly was the key.