4 months ago
Issue: Intermittent service unavailability causing connection failures and timeouts.
Symptoms:
Connection timeouts: ERR_CONNECTION_TIMED_OUT when accessing /api/admin/auth/verify-password
DNS resolution failures: ERR_NAME_NOT_RESOLVED for the service hostname
Intermittent availability: Service failed at 7:04am and 7:55am, then recovered at 8:01am
Evidence:
When the service is up, API endpoints return expected HTTP responses (e.g., 401 for unauthenticated requests), confirming the hostname resolves and the server responds
During outages, requests fail with connection timeouts or DNS resolution errors
The ~6-minute recovery window (7:55am → 8:01am) suggests automatic restarts
Is this because of the hobby plan? AI agents suggest that hobby sleeps, but Railway docs say otherwise. I'm happy to pay more, I Just need my service working...! I'm missing registrations.
9 Replies
4 months ago
Hi there, Hobby Plan services doesn't sleep (unless serverless is enabled). I don't think your issue is related to your current plan.
4 months ago
How are you making the requests to that API? using the public url?
I host the front end on Vercel, so use external urls.
This morning my flow was an attempt at 7:00am, which resutled in no server log. Another at 7:55am which resulted in no server log. And again at 8:00am which worked.
4 months ago
No logs even on the HTTP logs tab?
Chrome shows logs like : Failed to load resource: net::ERR_CONNECTION_TIMED_OUTUnderstand this error
7:00am: baby-manager-production.up.railway.app/api/admin/auth/verify-password:1 Failed to load resource: net::ERR_CONNECTION_TIMED_OUT
7:55am: baby-manager-production.up.railway.app/api/admin/auth/verify-password:1 Failed to load resource: net::ERR_NAME_NOT_RESOLVED
Opus suggests "The errors you're seeing are classic symptoms of Railway cold starts combined with no retry logic on the client side."
4 months ago
Then I shall close this thread in favor of that one!
Status changed to Closed medim • 4 months ago