[Bug]: [Bug] Official Railway librechat-official template fails /health healthcheck on deployment

[ylnztrkk]

What happened?

Bug: Official Railway "librechat-official" template fails healthcheck

Environment

• Platform: Railway
• Deployment method: Official Railway "librechat-official" template
• Container image: "ghcr.io/danny-avila/librechat-dev:latest"
• Region: "sfo"
• Replicas: 1
• Healthcheck path: "/health"

What happened

I deployed LibreChat twice using the official Railway "librechat-official" template.

Both deployments failed in exactly the same way.

The supporting services were deployed successfully:

• MongoDB
• VectorDB
• RAG API
• Meilisearch

However, the LibreChat service failed during the Railway network healthcheck.

Deployment result

Initialization (00:00)
Pre-deploy (00:10)
Deploy (01:04)
Network › Healthcheck (04:51)
Healthcheck failure
Post-deploy: Not started

Railway reports:

Deployment failed during the network process
Healthcheck failure

The LibreChat service configuration shows:

Deployed via Docker Image:
ghcr.io/danny-avila/librechat-dev:latest

Healthcheck path:
/health

Restart policy:
on failure

Restart policy max retries:
10

The container is configured with:

HOST=0.0.0.0
RAILWAY_RUN_UID=0

I also tried adding "RAILWAY_RUN_UID=0" and redeploying, but the exact same healthcheck failure occurred.

Expected behavior

The official Railway template should deploy successfully and the LibreChat service should become healthy and respond successfully to:

/health

Actual behavior

The LibreChat service deploys but never passes Railway's healthcheck and fails after approximately five minutes.

This happened twice with the same official template.

Request

Could you please check whether the current "librechat-official" Railway template and its "latest" Docker image are compatible with the current Railway deployment configuration?

In particular, please check:

1. The Railway healthcheck configuration and "/health" endpoint.
2. The container port / "PORT" configuration.
3. The current "ghcr.io/danny-avila/librechat-dev:latest" image compatibility with the Railway template.
4. The Railway dynamic variables used by the template.
5. Whether the official Railway template needs to be updated.

Version Information

LibreChat: ghcr.io/danny-avila/librechat-dev:latest

Deployment platform: Railway

Railway template: librechat-official

Region: SFO

Healthcheck path: /health

The official Railway template was deployed twice and both deployments failed at the Network → Healthcheck stage.

Steps to Reproduce

1. Open the official Railway LibreChat deployment template:
   https://railway.com/new/template/librechat-official

2. Deploy the template on Railway.

3. Allow Railway to provision the required services:

   • MongoDB
   • VectorDB
   • RAG API
   • Meilisearch
   • LibreChat

4. The supporting services deploy successfully.

5. The LibreChat service starts deploying using:
   ghcr.io/danny-avila/librechat-dev:latest

6. Railway reaches the Network → Healthcheck stage.

7. The healthcheck uses /health.

8. After approximately 4 minutes and 51 seconds, the healthcheck fails and the deployment is marked as failed.

9. I repeated the deployment twice and received the same result both times.

10. I also added RAILWAY_RUN_UID=0 and redeployed, but the same healthcheck failure occurred.

What browsers are you seeing the problem on?

No response

Relevant log output

Railway deployment output:

Initialization (00:00)
Pre-deploy (00:10)
Deploy (01:04)
Network → Healthcheck (04:51)
Healthcheck failure
Post-deploy: Not started

Railway message:

"Deployment failed during the network process"

Deployment configuration:

Docker image:
ghcr.io/danny-avila/librechat-dev:latest

Healthcheck path:
/health

Restart policy:
on failure

Restart policy max retries:
10

The other services (MongoDB, VectorDB, RAG API and Meilisearch) deployed successfully, but the LibreChat service did not pass the healthcheck.

I do not have additional container error logs because Railway only shows the healthcheck failure in the deployment view.

Screenshots

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[berry-13]

Thanks for the detailed report. I tried the official librechat-official template on Railway. Couldn't reproduce the issue. The LibreChat service passed the /health healthcheck both times.

• EU West: fresh deploy went Online in about 90 seconds. Server was listening on port 8080 about 13 seconds after container start.
• US West: moved the whole stack (LibreChat, MongoDB, VectorDB, RAG API, Meilisearch) to US West. Passed healthcheck again. GET /health returned 200 OK.

Both runs used ghcr.io/danny-avila/librechat-dev:latest with default variables. The port, HOST=0.0.0.0 and healthcheck path all look fine. RAILWAY_RUN_UID=0 isn't needed.

To narrow down what is different in your deployment, could you share:

1. LibreChat service Deploy Logs from a failed attempt. Open the failed deployment in Railway, then the Deploy Logs tab. A healthy startup ends with Server listening on all interfaces at port .... If that line is missing, the lines before it will show where startup stops.
2. Whether you changed any template variables before deploying. In particular CONFIG_PATH, MONGO_URI, or anything added yourself.
3. Whether MongoDB showed as Online when the LibreChat healthcheck started.

If redeploying LibreChat on its own works now, that points to something transient on Railway's side at the time.

[ylnztrkk]

Thanks for checking.

One additional detail from my side:

I am using the Railway Trial plan.

After changing the deployment region to Netherlands, the original deployment issue appeared to be resolved and LibreChat initially started successfully.

However, LibreChat has now started crashing repeatedly. Restarting the service only brings it back temporarily before it crashes again.

I checked the logs, but I cannot clearly identify the cause of the current crashes. I still see the previous disk-space error in the logs:

Post-listen initialization failed: available disk space of 231440384 bytes is less than required minimum of 524288000

However, I am not sure when this log entry was generated because the timestamps in the Railway logs appear to be incorrect, so I cannot confirm that this disk-space error is the cause of the current crashes.

At this point, the only clear observation I can make is:

• SFO: deployment failed with the disk-space/healthcheck issue.
• Netherlands: the deployment initially started successfully.
• Netherlands: LibreChat is now repeatedly crashing/restarting.

Since I am using the Railway Trial plan, could the Trial environment/resources also be considered when investigating the repeated crashes?

I will continue checking the Deploy Logs for a more specific error.