a year ago
Testing out Railway in hopes to move a Nextjs app here from Vercel.
Dockerized Nextjs runs fine, but when I point the production domain to the Railway app, seeing it's intermittently crashing and restarting with no logs (roughly a few times an hour). Unable to reproduce running container locally so far.Any ideas?
8 Replies
a year ago
The only log related to the crash is "container event container died". But, not seeing anything related to the actual reason (unless I am missing something).
Attachments
a year ago
Those would be runtime event logs from Railway.What do your memory metrics look like?
a year ago
Here is CPU and memory metrics when the app was in production.Reviewing the logs closer, looks like it actually crashed ~30 times during this 1.5hr period.
Attachments
a year ago
When using nixpack instead of a custom Dockerfile, it still crashes but return:"ELIFECYCLE Command failed with exit code 137."Leading me to believe a RAM issue. Will try upgrading and see if it goes away
a year ago
That's why I was thinking, an out of memory issue, the metrics recording may not capture very fast spikes due to the recording frequency.
After upgrading to Hobby make sure to do a redeploy.
Status changed to Solved railway[bot] • about 1 year ago
a year ago
After upgrading, so far so good!