cockpit
cockpit copied to clipboard
cockpit-project
Add tab in services for software that has been restarted.
Page:
It's hard to track down something that is set to auto restart on error, if there was an option to view servers that have been restarted or last loaded it would make it easy to track down issues.
I have something that keeps restarting but I don't know what it is.
We could show in the listing also how long each service was running and then enable sorting by this column. While these are reasonably easy to implement (we already have this data in details) I am mostly concern if it would fit into the UI and if it would not be too disruptive in most of the cases.
If there's some way to check to see if the process was started due to an auto restart on failure, we should show a warning and filter to these warnings to the top (similar to errors, but sandwiched between errors and the rest).
It would be a warning because it might be doing things as expected and perhaps hits some corner case crasher eventually and is restarted and processing things correctly.
If we cannot determine restart errors directly with systemd, we could determine it for a subset (where it's a quickly reloading service, like when it depends on a local or network filesystem that isn't mounted for some reason):
What if we some kind of warning for services that auto restart, are currently running, and have been running for < 10 seconds (or some other short timeframe)?
This could make it so Cockpit does the work to track down these problems without people having to do extra work.
I've hit this kind of issue before, not just for services, but also containers. (If we handle something like I suggested above here, we should probably also add a similar feature to Cockpit-Podman.)
What if we some kind of warning for services that auto restart, are currently running, and have been running for < 10 seconds (or some other short timeframe)?
This sounds like it could work?