Hey there! We've found the following might help you get unblocked faster:

• 🧵 Healthchecks Failing in US East and US West

• 🧵 Why healthcheck always fail

• 🧵 Healthcheck "service unavailable" but endpoint works when deployed

• 📚 Configure Healthchecks

If you find the answer from one of these, please let us know by solving the thread!

what’s the health check config?

Uploaded a screenshot of what I've tried in the UI. Obviously it's purple because I didn't actually deploy it since it was just for the sake of the screenshot. I've tried /health and /health/

I've also tried to configure it on my railway.toml

[deploy]
startCommand = "./start.sh"
restartPolicyType = "ON_FAILURE"
restartPolicyMaxRetries = 10
healthcheckPath = "/health/"
healthcheckTimeout = 300

Try removing the railway config file, and instead only use railway ui

I tried UI only many many many times before ever attempting via the railway.toml. Neither work.

Oh do you mean do not use the railway.toml any of the configs? I can try that

Im pretty sure the health check endpoint should only output “OK”, try making it output that instead of html

I can build an endpoint that returns a 200 and OK, but the docs don't say that is specifically needed. If that works, maybe I can submit a doc update request.

I am also going to test UI-only config without my railway.toml.

I built an endpoint, /railway-deploy-health/ , that returns an empty 200:

@require_GET
def railway_deploy_health(request):
    """Simple healthcheck endpoint for Railway deployment that returns only 200 OK."""
    return HttpResponse(status=200)

 path("railway-deploy-health/", railway_deploy_health, name="railway_deploy_health"),

Updated my railway.toml with that endpoint as the healthcheckPath:

[build] builder = "DOCKERFILE" [deploy] startCommand = "./start.sh" restartPolicyType = "ON_FAILURE" restartPolicyMaxRetries = 10 healthcheckPath = "/railway-deploy-health/" healthcheckTimeout = 300 [services.web] source = "." builder = "DOCKERFILE" startCommand = "./start.sh" # Health check configuration - waits for django-health-check to return 200 # Timeout allows for database migrations and service startup healthcheckPath = "/railway-deploy-health/" healthcheckTimeout = 300 restartPolicyType = "ON_FAILURE" restartPolicyMaxRetries = 10 [services.worker] source = "." builder = "DOCKERFILE" startCommand = "celery -A config.celery_app worker -l info --concurrency=2" [services.beat] source = "." builder = "DOCKERFILE" startCommand = "celery -A config.celery_app beat -l info"

The build completes and starts the health check, but fails:

2025-09-03T00:38:17.281512726Z [inf] importing to docker

2025-09-03T00:38:29.576556038Z [inf] importing to docker

2025-09-03T00:38:38.596309450Z [inf] Build time: 29.32 seconds

2025-09-03T00:38:59.913306491Z [inf]

2025-09-03T00:38:59.913351633Z [inf] ====================

2025-09-03T00:38:59.913356668Z [inf] Starting Healthcheck

2025-09-03T00:38:59.913361166Z [inf] ====================

2025-09-03T00:38:59.913365678Z [inf] 

2025-09-03T00:38:59.913369348Z [inf] Path: /railway-deploy-health/

2025-09-03T00:38:59.913373028Z [inf] Retry window: 5m0s

2025-09-03T00:38:59.913378730Z [inf]

2025-09-03T00:39:09.987477679Z [inf] Attempt #1 failed with service unavailable. Continuing to retry for 4m49s

2025-09-03T00:39:20.024190309Z [inf] Attempt #2 failed with service unavailable. Continuing to retry for 4m39s

2025-09-03T00:39:22.272263749Z [inf] Attempt #3 failed with service unavailable. Continuing to retry for 4m37s

2025-09-03T00:39:26.371656359Z [inf] Attempt #4 failed with service unavailable. Continuing to retry for 4m33s

2025-09-03T00:39:34.418375267Z [inf] Attempt #5 failed with service unavailable. Continuing to retry for 4m25s

2025-09-03T00:39:50.502391870Z [inf] Attempt #6 failed with service unavailable. Continuing to retry for 4m9s

2025-09-03T00:40:20.615050397Z [inf] Attempt #7 failed with service unavailable. Continuing to retry for 3m39s

Despite the server running:

2025-09-03T00:38:58.000000000Z [inf] Starting Container

2025-09-03T00:39:16.371057708Z [inf]

2025-09-03T00:39:16.371064754Z [inf] === Starting Gunicorn ===

2025-09-03T00:39:16.371070355Z [inf] Binding to: 0.0.0.0:8080

2025-09-03T00:39:16.933354939Z [err] [2025-09-03 00:39:16 +0000] [1] [INFO] Starting gunicorn 23.0.0

2025-09-03T00:39:16.933361698Z [err] [2025-09-03 00:39:16 +0000] [1] [INFO] Listening at: http://0.0.0.0:8080 (1)

2025-09-03T00:39:16.933366101Z [err] [2025-09-03 00:39:16 +0000] [1] [INFO] Using worker: sync

2025-09-03T00:39:16.942190005Z [err] [2025-09-03 00:39:16 +0000] [51] [INFO] Booting worker with pid: 51

2025-09-03T00:39:17.022531808Z [err] [2025-09-03 00:39:17 +0000] [52] [INFO] Booting worker with pid: 52

In case it's helpful, this is how I have my healthchecks configured, in my railway.json file:

  "deploy": {
    "healthcheckPath": "/healthz",
    "healthcheckTimeout": 60,

…and this is the code that responds to requests there:

if (req.url === "/healthz") {
      res.writeHead(200, { "Content-Type": "text/plain" });
      res.end("OK");
    }

https://stackoverflow.com/questions/43380939/where-does-the-convention-of-using-healthz-for-application-health-checks-come-f

"/healthz" is a 404
www.rosternerd.com/healthz

I don't have /healthz set up so you're going to get a 404. My endpoint is currently /railway-deploy-health. I can set up /healthz if it's a requirement, but I don't see if the docs where it says that's what it has to be called.

I got this working. The issue was that my project had the Django setting SECURE_SSL_REDIRECT set to true by default. That being the case, the health check's HTTP requests were receiving 302 redirect codes as my app was redirecting to HTTPS. Once I disabled SECURE_SSL_REDIRECT the health check started working.

For posterity: After removing the SECURE_SSL_REDIRECT setting, all of the health check endpoints that I set up started working. That means that a 200 OK is not required. It also means that HTML is fine.

I appreciate all of the helpful suggestions!

To close the loop on /healthz - no it's not required to do it this way, it's just a convention that Google uses.

Thanks for the update. I did read the link you provided for that additional context. I'm actually going to keep the /healthz endpoint for railway to use for these health checks and keep it separate from my other /health endpoint which has more going on than is needed for these railway checks. Thanks again!

SECURE_REDIRECT_EXEMPT = [r'^healthz/$'] 

This has solved the issue for me. So you dont have to deactivate it completly.