Healtheck error and logs broken

[warg]

Hello,

LibreTranslate is running for some time without issues. In the logs I see that the HealthChecks died for unknown reasons:

Furthermore clicking "Download Full Logs" downloads a text file with the exact same (very few) log entries but throws an error that the log file wasn't found.

Looks like Cloudron v7.5.0 broke it or it's unstable.

Best Regards,

[girish]

Not seeing this in both our instances. I tired a test install and that doesn't have this error either. Anyone else seeing this?

Do you see this error even after app restart?

[warg]

Restarting the app did help to get the log logging again but downloading the full log file is still broken.

Looks like there's an error in the browser console of Firefox:

[girish]

Do you get the logs error with other apps as well? Just wondering if this is a general log viewer issue in 7.5.0 . Maybe we introduced a regression.

[warg]

Health Checks just died again so looks like a reoccuring issue without any user-related activity.

Downloading full logs (error "Log file(s) not found") is broken for other apps as well, e. g. NextCloud. So we have 1 app-specific issue and 1 app-unrelated issue (regression).

[nebulon]

for the log file issue, can you enable remote SSH support for us and send a mail to support@cloudron.io with your dashboard domain so we can debug this directly?

[warg]

I have sent an email (actually 2 because I forgot details in the first one). Thank you.

[nebulon]

Thanks for providing this. The issue with downloading full logs is indeed a regression in 7.5 and is fixe in the upcoming release then.

For Libretranslate, it seems the app as such is running, but still shows those timeouts in the logs for the healthcheck. Not sure yet what the root cause for this is.

[warg]

@nebulon said in Healtheck error and logs broken:

For Libretranslate, it seems the app as such is running, but still shows those timeouts in the logs for the healthcheck. Not sure yet what the root cause for this is.

Looks like a regression too because it didn't happen with previous Cloudron version and there was no app update.

[nebulon]

Actually the timeouts for the healthcheck are quite infrequent (about faililng twice per hour, while we perform the check every 10 seconds!) and seem to be due to temporary server load. I don't think there is anything to be fixed as such besides adding more hardware resources. So in the end there is likely nothing to be concerned as long as the app itself is working.

[warg]

How comes that you think the hardware resources is an issue? This sounds very unlikely to me (stats for last 24h):

• RAM is 3 of 16 GB used
• CPU is <= 20 % exhausted
• disk is nearly not even half full + high read/write rates

If it's a performance issue, this is caused on Cloudron's side somewhere. Furthermore there is zero user activity on my the machine nor does the timeouts happen at times where something is done, e. g. when I'm not even connected to the machine and no updates run.

Also it looks like a bug that the app is reported as online/running when healthcheck fails.

[nebulon]

As said, the healthcheck succeeds nearly all the time, so besides the log line there is nothing to worry about.

[warg]

I see. So basically the healtcheck logging in the LibreTranslate app is different to the other apps and thus I wasn't aware the healthcheck does succeed most of the time because in the log you see no information about successful healthchecks, only failed ones it looks like while other apps show both.

[nebulon]

Correct, the logging of the upstream apps as such is app specific, so in this case only failed requests are show.

I guess we can close this thread then as resolved?

[warg]

Sounds good, let's close it. Thank you.