supervisor container created at unexpected times

Hello,
I have noticed something interesting. The supervisor container in my balena-cloud does not seem to follow the same creation patterns as the other containers, as it seems to “die” at unexpected times. For example, I see my other containers have been up and running for about 4 days. While my own services report a created_at at “4 days ago”, the supervisor seems to have been about once a day. Is this normal? Is the supervisor expected to keep running just like any other, or are there some up/down pattern that are not mentioned in the docs?

Thank you!

Hi,
can you please have a look at the supervisor logs with journalctl -r -u resin-supervisor and check for errors? You should be able to identify the supervisor restart and the log messages with the cause should be somewhere around.

@mtoman

@mtoman interesting to noticed that the logs only being at 14:26 today, when the other containers have been running for 4 days…

Thanks for coming back, do the other containers have logs older than 14:26? We do not store logs by default because of SD card wear so you can only see the latest logs. This greatly depends on how much logs your application produces. You can enable persistent loggin via the balenaCloud dashboard in the Device configuration tab and having a look after a next restart. Could you please also share which device and what balenaOS version you are using?

@mtoman a storm just hit the factory so the electricity went off and the state of the app got lost. I will keep an eye on the issue and post here again as soon as I can reproduce it. in the meantime, thanks a lot!