PLEASE HELP: Coolify Services/Server Fails after 1-2 days and only server reboot resolves

[tieandrews]

TLDR: Setup Coolify running on Ubuntu desktop and having coolify server/services constantly fail every 1-2 days, reboot/restart resolves, trouble finding culprit

So as the TLDR implies, I have a Ubuntu 24 desktop on which I want to try to run Coolify and services within as it's hardware I own and don't want to pay for another hosting service.

I've setup SSL/HTTPS for the services deployed (Uptime Kuma, OpenWebUI, etc.) and running through cloudflare tunnels using cloudflared running via CLI (not in the built in Coolify Claoudflared service as it sounds like this may have more issues?), it works well for a day or two but then either Coolify views the local server as "not reachable"/unhealthy, and all the services show "no server available" errors.

I've setup local scripts checking/logging cloudflared status and no clear correlation between when the cloudflared service fails and the general service failures. Also the desktop is connecting via Wifi, and I know it's a potential failure point, so I set up a script to check for wifi access failures every 60s and again no clear correlation between when there are wifi disconnects and the Coolify services fail/disconnect or the cloudflared tunnels fail.

A reboot of the machine running the services typically resolves the issues for 1-2 days, but after that time frame they go down again.

I am looking for any tips for debugging or identifying failure points on coolify services so I can start to rely on them for moderate usage outside of personal use. I know there are a lot of details buried in there so happy to provide more info but just annoyed trying to keep things up and running reliably. Thanks in advance!

System Info:

• local server: Ubuntu Desktop 24.04
• coolify version: v4.0.0-beta.442
• Traefik configured to use SSL certs for https as per the user guide
• RAM: 32gb

Can provide any other system info that might be helpful in debugging. TIA!

[maietta]

How was Docker installed?

Was is a snap package? Official manual install, or convenience script?

[tieandrews]

Thanks @maietta , Docker was installed with the Docker Desktop setup for Ubuntu, confirmed manually that snap wasn't used with:

snap list docker

Nothing returned.

[djsisson]

You mention wifi, i assume you have fixed the ip

You can check coolify logs or journalctl at start of downtime to see whats gone wrong

[tieandrews]

Thanks @djsisson , yes I've pinned my machines IP address, didn't do anything at the router though as access to it is not easy at the moment. Not sure if there are other networking things to consider with this sort of setup.

I'm unclear if it's an issue that I'm running both coolify, and ~5-6 services on the same machine vs coolify on the local machine then services on a remote server.

This is the Uptime Kuma showing how often the services are crashing, recovery is me rebooting the machine, never self recovers even with everything setup to restart: unless-stopped.

I can't figure out how to inspect the coolify server logs from around/before disconnections from the cloudflared tunnel, any tips?

[djsisson]

My suspicion is a memory issue, as the daemon will kill any container if you run oom. Or some kind of hardware issue.

Check journalctl -u docker.service or any docker logs xxxxx where xxxxx is the container like coolify

After a restart run htop and check resource usage.

Also you mention running cloudflared via cli not via docker, can you check the logs for that too, do you have autoupdate turned on?

[Cainor]

Facing same issue .. mine is every about 12 hours, the container just hangs. I was able to use the healthcheck to see if a container is unhealthy. But I want to automate restarting or spawn new containers to keep serving.