Railway favicon

Railway outage on 2026-10-28

railway.com
October 2026 28 Minor incident

Railway's build and deploy pipeline broke three separate times on 28 October, starting at 10:07 UTC and running through to roughly 20:21 UTC. All three incidents sat on Railway's side of the system, in build infrastructure and deployment initialisation, not in your code.

Started

10:07 UTC

Duration

Lasted 37m

Source

IsDown

Next time

Railway is degraded. Is your product working?

An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.

50 monitors for your own websites. No card.

What happened?

The first incident started at 10:07 UTC, when Railway identified a problem with build infrastructure in US West causing builds to fail before any build steps ran, with the error "build daemon returned an error". Railway said running services were not affected and users outside US West were not affected, and suggested redeploying as a workaround. By 10:36 UTC Railway said it had applied a fix and was monitoring recovery. That incident ran 37 minutes. A second and separate incident began at 14:56 UTC, this time hitting builds intermittently across both EU West and US East. Railway's update said the cause had been identified and that it was working with upstream providers, again noting running services were not affected. No further update was filed after the 14:56 note, so the 2h 21m duration on record is the longest of the day, and it closed without a public confirmation message of what the fix was. A third incident started at 17:49 UTC, this time about deployments getting stuck in an "initializing" state rather than builds failing outright. Railway said the cause had been identified and it was "actively fixing," but, as with the second incident, no closing update appears in the record. That incident lasted 1h 32m. Across all three, Railway consistently said running services already deployed were not affected, meaning the damage was confined to anyone trying to ship new code or redeploy during those windows. No user reports were filed in this window. Railway's own history shows 35 incidents in the last 90 days with a median duration of 1 hour 30 minutes, so a day with three separate build and deploy problems is not unusual for this service.

Learning

Builds and deploy pipelines are a different layer from the running service itself, and that split is exactly what this day shows: Railway said three separate times that already-running services were untouched, while the ability to ship new code was broken for stretches adding up to several hours. A ping to your live application's URL would have stayed green through all of this, because the application was never down, your deploy pipeline was. That is the general shape to watch for with any platform like this: uptime checks against your endpoints tell you about the thing already running, not about whether you can change it. Over the last 30 days UptimeRobot's own checks logged 22 incidents against this provider with a median length of 1h 22m, which lines up with what Railway itself reports, so the pattern holds across more than one day. If your team only watches its own application and not Railway's platform status, a day like this looks like silence until someone tries to deploy. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.

4.7
stars out of 5
284+ reviews on

Start monitoring in 30 seconds.

There's nothing to install. No credit card required. 50 monitors for free.