- Level
- confirmed
- Worst impact
- Degraded
- Components
- US East (Virginia, USA), US West (California, USA), TCP Proxy (Metal), EU West (Amsterdam, Netherlands), Southeast Asia (Singapore)
- Regions
- apac, eu, us
- Down Index
- 33of 100
What to do
Check whether you use Railway US East (Virginia, USA), US West (California, USA), TCP Proxy (Metal), EU West (Amsterdam, Netherlands) and 1 more in apac, eu, us. Check your stack
Paste a package.json, lockfile, Terraform or a list of domains at down.guide/stack to see every provider you depend on.Confirm your retries, queues and scheduled jobs caught up after the 5h 16m it lasted.
Anything that failed between 2026-03-30 12:20 UTC and 2026-03-30 17:37 UTC may need a replay.Subscribe to Railway's feed to hear about the next one first. RSS feed
https://down.guide/feed?provider=railway works in Slack (/feed subscribe), Discord feed bots and any RSS reader.
Updates from Railway 4
- investigating
While updating our edge CDN to allow us to identify cached assets on a per-domain basis, we enabled a feature called “Surrogate Keys”. Unknowingly to us, this bypassed our “if CDN is disabled for this domain, pass directly to origin” behavior, which meant that we accidentally enabled CDN caching capabilities for domains that didn’t previously have it enabled. Because of that, the following assets may have been cached during this 30 minute period even with domains with the CDN feature disabled: Any content with public Cache-Control headers Images, videos and other known-static assets No HTML would’ve been cached; this is always considered dynamic At 11:30 UTC, once realizing the mistake, we immediately pushed an update which reverted this behavior and purged the cache of all assets globally.
provider
- monitoring
This is a retroactive incident that was resolved at 11:30 UTC. We are still determining impact.
provider
- monitoring
We are seeing reports that even content without cache-control set was erroneously cached. We are still determining impact.
provider
- resolved
This incident has been resolved. A postmortem is available at https://blog.railway.com/p/incident-report-march-30-2026-accidental-cdn-caching.
provider
Reports count in aggregate; nothing about you is shown.
Corrections
None so far.
Something wrong on this page? Report an error or send us a note. Every fix is listed here.
Questions about DWN-b6b9f3547b
Is Railway still down?
No, this incident was resolved 2026-03-30 17:37 UTC. For Railway's live state, see https://down.guide/railway.
How long did it last?
5h 16m, from 2026-03-30 12:20 UTC to 2026-03-30 17:37 UTC, by the times Railway posted.
What was affected?
Railway named US East (Virginia, USA), US West (California, USA), TCP Proxy (Metal), EU West (Amsterdam, Netherlands) and 1 more in apac, eu, us.
What caused it?
Railway hasn't published a root cause. If it does, it shows here.
Where does this come from?
Railway's official status page, https://status.railway.com, quoted verbatim. down.guide's own probes, reports and social counts are labeled as ours.