Did you hit your size limit(5GB)(check here ...Volume -> Metrics)
Try setting RAILWAY_RUN_UID=0 env variable in your service
more info : https://docs.railway.com/volumes/reference#caveats

Size limit for the volume is not exceeded (300MB used from 500MB volume). I forgot to mention that it was working perfectly for a few weeks. Then it suddenly crashed a few days ago, could not make it work since.

RAILWAY_RUN_ID=0 was not set before. Added it now but did not help, crashes the same way as before being set.

its RAILWAY_RUN_UID=0 ..not just ID

Also check this https://station.railway.com/questions/mongo-db-not-working-ad53088c

Right, it's a typo in my previous reply. Setting RAILWAY_RUN_UID=0 doesn't change anything. Also, as I mentioned, it was working without it for weeks.

Thanks. It's a vague description of the issue, but it doesn't seem the same issue I am having. I can execute Restart or Redeploy on the instance, but it doesn't start up. As I can tell from the logs the instance is launched, but start up fails with an ungraceful stop while attempting recovery from the initial crash.

I figured out the issue by SSH-ing into a deployment where the volume is mounted.

Information about the volume usage on the Metrics tab is wrong. Metrics showed ~300MB from 500MB used, but in fact there was no free space at all on the volume, space was 100% used up. Resizing the volume to 1GB helped, MongoDB now sucessfully starts.

However, Metrics still show volume usage ~300MB from 1GB, altough ~424MB is really used. (Formatted 500MB volume gives that much usable space, 1GB formatted volume gives ~869MB of available storage space.)