The log shouldn’t be clearing just because it crashed, I can still get to the last logs of my stopped add-ons without issue. They should only clear when the add-on starts up again.
Do you have “watchdog” turned on? If so you need to turn that off. That way next time it crashes you can go check out what happened then start it up manually.
I assume you don’t have log_level set to suppress almost everything right? Make sure you have log_level set to debug or at least info in the configuration.
Other then that I’m not really sure, logs show up there for me. I don’t have a lot of Node RED crashes but had some when trying to leverage a custom npm package incorrectly and I was able to debug it with the logs.
I notice you have an “app configuration” option in the bottom left of your navigation which i only see when I use the mobile app but that looks like too big a screen for the mobile app. What app are you using there? Maybe try going to the website with a browser to see if the app has a problem with the something about that screen?
Another option is you could try the portainer add-on. If you browse the list of containers in there you can find the node red one and see its logs by clicking this icon:
I can do that even for stopped containers so it should still show up. But for me that shows me the same thing as I see in that tab so I’m not sure that will help in your case.
Is the error occurring during start-up by the way or does it start running fine and then crash at some point?
I did get a fatal level message but no critical one and then it shut down. I would lower it to at least fatal if not error. Maybe try fatal first but you might only get one message that way, if its not informative enough you’ll probably need to drop another level.
Logging is set to info. I also have set up portainer. However, that revealed that not just the addon is stopped but the host is rebooting. I will monitor this for the next hours, to see if this is caused by the addon or the host itself is instable.
Oh that is very different. And yea that does probably clear out the logs.
In retrospect, I should’ve asked about this more. That sounds very wrong and like something that definitely shouldn’t happen. Debugging what is wrong in HA OS itself is a bit beyond me if that’s the situation you’re in. Fortunately that is definitely a supported install so hopefully one of the devs can help you out.
I might suggest taking this to the homeassistant_os channel in the HA discord. That’s a more direct support channel then the forum and I think you might need that in this case.