5 hours ago
Hi Railway team,
We received a "Server Outage Affecting Your Service" notice for our Postgres
service (hardware failure, "Expected Resolution: Within 30 minutes"), project:
magic-coloring, environment: production.
The service is still completely unreachable well past that window:
-
2026-09-07T02:02:38Z — first connection failures from our worker service:
Error: Connection terminated due to connection timeout
at /app/node_modules/pg-pool/index.js:45:11
at async connectDB (/app/config/database.js:89:3)
(pg client with connectionTimeoutMillis = 10s — the TCP connection gets no
response at all, it is not an auth or query error)
-
The same timeout repeats on every attempt since. Our worker runs on a 5-minute
cron schedule, so we have a continuous record of failed attempts up to now.
-
Our web service connects to the same Postgres instance and is fully down —
the site cannot serve any request.
Could you please:
-
Confirm this is the same hardware-failure incident and share an updated ETA.
-
Confirm whether the Postgres volume/data is intact and will come back as-is,
or whether any snapshot restore is involved (we need to know about data
integrity for our customers).
-
Tell us if anything is required from our side.
Project: magic-coloring — https://magic-coloring-production.up.railway.app/
Affected: Postgres (production); web + worker services depend on it.
Thank you!
1 Replies
5 hours ago
The host maintenance event for your Postgres service has resolved, meaning the underlying recovery is complete, but occasionally a service does not come back on its own afterward. Your Postgres volume is attached and intact (217 MB of data on it), so no data was lost. To bring Postgres back, open the Postgres service, press Cmd+K (Ctrl+K on Windows/Linux) to open the command palette, and run "Redeploy source image." This will not delete or reset your data. Once Postgres is accepting connections again, your web and worker services should recover as well (if the web service stays crashed, redeploy it the same way using "Deploy latest commit"). For mission-critical workloads going forward, we advise implementing a high-availability configuration so a single-host failure no longer takes production down.
Status changed to Awaiting User Response Railway • about 5 hours ago