Crashed service still working
Hi,
I received an email telling me my service crashed. When I go to the dashboard the deployment said it's crashed, however, everything is alright, and the api is serving requests succesfully. I would expect it recovered/restarted itself, but the deployment is still marked as crashed somehow?
34849d22-d685-4e73-8858-fbd4fe42ea65
11 Replies
Project ID:
34849d22-d685-4e73-8858-fbd4fe42ea65!t
New reply sent from Help Station thread:
This thread has been escalated to the Railway team.You're seeing this because this thread has been automatically linked to the Help Station thread. New reply sent from Help Station thread:
Sorry bout this - we're going to investigate why this was marked as crashed when it's evidently running.As a workaround, can you try clicking "Restart" on the deployment?You're seeing this because this thread has been automatically linked to the Help Station thread.
"Restart succesfully" in the bottom corner notification
"Crashed x seconds ago..." on the service canvas
So service still working but still says "crashed". Logs are normal with requests going on. The only weird log is "container event container died" that would explain the "crashed"
Got an "container event container restart" log and now it's green and marked as "completed"/"deployment successful"
just fyi, it's not the first time it happens, I got emails on
- 19.07.2024
- 22.07.2024
- 23.07.2024
but since everything is running smoothly on my side, I don't really bother pressing "restart" so the color is green instead of red π
Do you have anything that's restarting the service at 8am/4pm UTC? We're seeing "restart" API calls coming in during those time
maybe they are using my cron template, it calls the API with a restart mutation
Correct I have 2 restarts: 8am UTC and UTC-8. Iβm using the graphql mutation straight away.
Yup, we found an issue where we don't pipe the correct events to our backend when a direct restart happens, so it registers as "crashed" in the dashboard.
nice, great news!
I got the "crashed" status at 8:00 again, so I believe it's not fixed yet, just fyi
to be fair they said found the issue not fixed the issue π
yeah that's true, I'll be a bit more patient ahah