6 hours ago
Project: github-runners (f9c38ce1-16fc-4377-8879-a3efaeeea398) Environment: production (62147b86-e2b2-4f62-aefc-cb518b1d78df) Service: runner-clean (f255c69b-0022-4b6b-af73-01e1a08e91d1), 15 replicas, us-west2, restart policy ALWAYS Deployment: af5d7521-5eb2-46fa-b468-956934a08859
Affected replicas
da7fae1a-bd4f-4930-9b48-bb24a6fca8d1: exit code 0 at 2026-11-17 21:15:16 UTC
77e7bf6d-691e-4696-9e62-3b9924f68f2d: exit code 0 at 2026-11-17 22:16:09 UTC
Observed
After the exit, neither replica produced any log line. Our startup script prints runner-start: reset complete first, and it never appeared.
Neither made any network or DNS calls afterwards.
Both had restarted correctly many times before. Their final exits look identical to the earlier ones.
The dashboard kept showing both as running.
By 18 Nov we were down to about 10 working replicas out of 15, until we redeployed.
Ruled out
The start script prints RUNNER-START-REFUSED and exits 1 on any failure. Neither replica logged it.
Needed from Railway
Exit status, restart count and scheduling events for both instances.
Why ALWAYS did not restart them.
Whether other replicas were affected this week.
0 Replies
Status changed to Awaiting Railway Response Railway • about 6 hours ago