a month ago
Service: relay (EU West, Amsterdam)
Deployment: 2d05e7d6
Builder: builder-ervuqb
Since roughly 08:58 BST this morning, builds fail at the snapshot fetch stage and never reach the build itself. No dependency install, no compile — nothing from the repository is ever read. The log is a bare retry loop roughly every 30 seconds:
08:58:38 failed to fetch snapshot
08:59:24 failed to fetch snapshot
09:00:25 failed to fetch snapshot
09:02:56 failed to fetch snapshot
09:05:06 scheduling build on Metal builder "builder-ervuqb"
09:05:12 failed to fetch snapshot
09:06:28 failed to fetch snapshot
The deployment sat in BUILDING for over 40 minutes without ever leaving the fetch stage.
Already tried: cancelled and redeployed several times, and pushed a fresh commit to trigger a new build. Identical failure each time — same stage, same message.
Ruled out: no Dockerfile or build configuration changes; the commit that triggered this touched only a JSON documentation file; previous builds on this service completed normally. The failure occurs before any repository content is read, so I don't think anything on my side can be causing it.
I can't test this by moving the service to another region — our outbound IP is allowlisted on a client firewall for an upstream API, so a region change would break that integration.
Two questions:
- Can someone check builder-ervuqb, or reschedule this service onto a different builder within EU West?
- Is anyone else in EU West seeing snapshot fetch failures today? The status page shows EU West builds as operational, so I'm trying to work out whether this is a fleet issue or isolated to my project.
3 Replies
Status changed to Awaiting Railway Response Railway • about 2 months ago
a month ago
This is in after an hour of replicated build errors
This deployment failed because Railway's build system could not fetch the code snapshot needed to start the build, not because of anything in your app or config. The build was rescheduled across five different builder machines over more than an hour, and each attempt failed immediately with a "failed to fetch snapshot" error before any install or compile step ran.
a month ago
And it just fixed itself for no apparent reason
a month ago
The failing attempts were all redeploys of one deployment, the build that first failed yesterday evening. That build failed before its code snapshot was saved, so every redeploy of it retried a download that could never succeed, no matter which builder picked it up. The builders in EU West are healthy and there was no regional snapshot issue today. Between the first failure and 10:19 BST this morning, no fresh-commit build reached the platform, only redeploys of that one deployment. The recovery was not random: a fresh commit pushed at 10:19 BST created a new snapshot and built normally in under a minute, and every push since has done the same, with nothing changing on our side in between. To avoid a repeat, deploy from a fresh push or redeploy a previously successful deployment. Redeploying a deployment that failed at the very start of its build will loop the same way. The service is live and healthy now.
Status changed to Awaiting User Response Railway • about 2 months ago
Status changed to Solved dizzydes90 • about 2 months ago