Was the JSON data stored in a dedicated volume or just within the container?

It was stored in a dedicated volume. The volume is named 'bodasofhiyalex-volume' and I had it correctly mounted to the path /app/data from day one. However, after today's US West outage and a subsequent container restart, the volume appears to be completely empty or corrupted. The filesystem mounted at /app/data no longer has my invitations.json file. Could the outage or the redeploy have wiped the persistent volume data?

Was the file stored under /app/data though? (Just a sanity check)

Yes! The code uses path.join(process.cwd(), 'data', 'invitations.json'). Since the working directory in Railway's container is /app, the absolute path where the file is written is exactly /app/data/invitations.json.

I've confirmed the volume is correctly mounted (cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e to /app/data) and the volume ID hasn't changed — no duplicate was created.

I also checked my code and confirmed it does NOT reinitialize or overwrite the file on startup. The only code that touches invitations.json is a read-only rescue endpoint. So the file wasn't cleared by my application.

The empty file came back after the service recovered from the outage at 05:09 UTC. This points to a storage backend issue during the incident recovery.

Request for Railway team: Can you check the underlying block storage snapshot for bodasofhiyalex-volume (ID: cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e) from before the US West outage (before 2026-08-10 03:39 UTC)? My invitations.json data was there before the incident — it should be recoverable from the snapshot if it wasn't permanently lost.

I've confirmed through Railway Agent support chat:

Volume is correctly mounted (ID: cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e to /app/data)

My code does NOT reinitialize or create the file on startup (verified in codebase)

The file was there before the outage, came back empty after recovery at 05:09 UTC

This is a data loss issue from the platform outage. Railway team: Can you check the block storage snapshot for this volume from before the incident started (2026-08-10 03:39 UTC)? My invitations.json data should be recoverable.

Can a Railway person help me please? 😔 My wedding invitations got lost and I’m really worried 😟

I’d suggest maybe disconnecting the volume from your service and trying to use the filebrowser template to see if anything is present in the volume. For a guide, look at this template

Following the milo's suggestion, I checked the volume using the file browser console and confirmed:

Volume is correctly mounted (ID: cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e to /app/data)

My code does NOT reinitialize or create the file on startup (verified in codebase)

The file exists with timestamp 2026-08-09 22:22 — before the outage started at 03:39 UTC

The file is completely empty (0 bytes)

This proves the file was there with data, but the CONTENT was wiped during/after the outage recovery while the file metadata remained intact.

Railway team: Can you check the block storage snapshot for this volume from before the incident (before 2026-08-10 03:39 UTC)? My invitations.json data should be recoverable from the snapshot.

Please Railway employees help me :( This are my wedding invitees and I really need to recover this file 😭

Volume ID: cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e

File timestamp: 2026-08-09 22:22 (BEFORE outage)

Current state: 0 bytes (AFTER recovery)

Outage window: 2026-08-10 03:39-05:20 UTC

Request: Check block storage snapshots from before 03:39 UTC

I've completed a full investigation with Railway Agent and confirmed this is a platform storage issue, not user error.

Evidence:

Volume mount is correct: bodasofhiyalex-volume (ID: cfc8778c-fbf7-49ae-9705-23d0e9cfeb2e) mounted to /app/data

Verified through console file browser: invitations.json exists on the volume but is completely empty (0 bytes)

Critical timestamp: File metadata shows 2026-08-09 22:22 — 2 hours and 17 minutes BEFORE the US West outage started (03:39 UTC)

My application code does NOT reinitialize or overwrite this file on startup (verified in codebase)

This was the ONLY copy of guest RSVP data (not in git, not backed up elsewhere)

What happened:

The file existed on the volume with data before the outage. After Railway's service recovery at 05:09 UTC, the file still exists but the content was completely wiped. The file metadata timestamp remained unchanged, proving the file wasn't deleted, it was truncated/corrupted during the storage backend recovery process.

What I need from Railway:

The US West outage incident affected our block storage backend. Can the Railway team check if there are snapshots or backups of the bodasofhiyalex-volume volume from before 2026-08-10 03:39 UTC? My invitations.json data should be recoverable from a pre-incident snapshot.

This is unacceptable for production data. I was about to migrate to Vercel Postgres for exactly this reason — to avoid data loss when using persistent volumes on Railway.

We're really sorry, but the RSVP data isn't recoverable on our side. We checked the storage host directly: there are no snapshots of this volume, and no backup schedules were ever configured, so there's no earlier copy to restore from.

We did check the platform side thoroughly. The volume was never moved or recreated, and it's still on the same host it's been on since July. The incident itself was network congestion in US West, which made services in the region slow or unreachable, and hosts in the region were under heavy load during that window. One thing worth noting from your screenshot: the file's last-modified time is about two hours before the outage began and didn't change through recovery. That would be unusual if the content had been wiped during the incident, since emptying a file normally updates its timestamp.

We know how much this data mattered, and we're sorry we can't bring it back. Enabling volume backups on anything you need to keep will give you a restore point going forward.