Hi,

Every HTTP request to my service is being refused at the Railway edge with

429 "rate limited". No request reaches my application at all. TCP traffic to

my Postgres service on the same project works normally.

Account: dev@mobilityhubafrica.com

Workspace: Tech Dev's Projects (personal) — 03cd24a7-b2df-4c60-a8ec-50a1a70d89ac

Project: endearing-bravery — 12d744f6-b89c-4d16-ba40-e5b49b01754e

Environment: production — d6344613-67a0-4dc8-914a-f74669c091c1

Service: mba-backend — def9723c-579f-4276-85f2-05aafdbdb977

Domain: mba-backend-production.up.railway.app (target port 3000)

Trace IDs for two refused requests, ~21 hours apart and from different edges:

x-hikari-trace: jnb1.w0as — Fri, 31 Jul 2026 18:43:11 GMT

x-hikari-trace: bcn1.h9g5 — Sat, 01 Aug 2026 15:43:29 GMT

Response received for every request, including /health and paths that do not

exist in my app:

HTTP/2 429

content-type: text/plain; charset=utf-8

content-length: 12

server: railway-hikari

x-railway-edge: bcn1

rate limited

Why I believe this is not my application:

1. My app's own rate limiter returns JSON, not plain text, and always sets

   RateLimit-* headers. This response has neither.

2. /health in my app has no rate limiter and no dependencies. A request to a

   route that does not exist in my app returns the same 429, so routing never

   reaches my code.

3. The HTTP log stream for the service is completely empty — the proxy has

   never forwarded a request to the application.

4. The behaviour is identical across two different edge locations (jnb1, bcn1)

   and has persisted for more than 24 hours.

5. Postgres on the same project is reachable and serving normally over the TCP

   proxy (reseau.proxy.rlwy.net:20853), so only the HTTP edge is affected.

Likely relevant: until today my service had a misconfigured start command

(prisma migrate deploy, which exits immediately instead of starting a

server). The container therefore exited and was restarted repeatedly for

roughly a day. I suspect that restart loop tripped an automatic protection at

the edge.

That misconfiguration is now fixed. The current deployment

(b2bf0136-5051-435e-b72c-da05c975b446, SUCCESS at 2026-08-01T16:47Z) has:

startCommand npm start (node dist/index.js)

preDeployCommand npx prisma migrate deploy

healthcheckPath /health

The build completes cleanly and the pre-deploy migration step succeeds

("No pending migrations to apply", connected over postgres.railway.internal).

Could you please:

1. Confirm whether a rate limit, throttle or usage restriction is currently

   applied to this project, service or workspace, and clear it if the cause

   was the restart loop described above.

2. Confirm my plan/usage state is not the cause — if a trial or credit

   allowance has been exhausted, please say so directly and I will resolve it

   on my side.

Thanks,

Nana Yaw