4 days ago
Subject: Custom domain intermittently returns ERR_CONNECTION_TIMED_OUT
Project: thorough-beauty / production
Service: daring-manifestation
Custom domain: app.ionixautomation.online
Railway domain: daring-manifestation-production.up.railway.app
Issue:
My custom domain (app.ionixautomation.online) intermittently fails to load,
showing "ERR_CONNECTION_TIMED_OUT" in the browser after 60+ seconds of waiting.
This happens on multiple networks (mobile data + WiFi) and in incognito mode,
ruling out client-side caching.
What I've verified:
-
DNS records are correct: CNAME app -> kjonv5il.up.railway.app,
TXT _railway-verify.app matches Railway's expected value
-
Domain shows verified (green checkmark) in Railway's Networking tab
-
Serverless/sleep mode is disabled on the service
-
Container logs show healthy "200 OK" responses during the same time window
the custom domain was timing out
-
The auto-generated Railway domain (*.up.railway.app) also occasionally
failed to load during the same period, then started working again
without any redeploy or restart
This is an active production app for paying subscribers (launched this week),
so I'd appreciate help investigating whether this is an edge/proxy routing
issue on Railway's side for this custom domain.
Happy to share screenshots/timestamps of the timeout if helpful.
11 Replies
4 days ago
This thread has been opened as a bounty so the community can help solve it.
Status changed to Open Railway • 4 days ago
4 days ago
Thanks for checking! I tried connecting via Turbo VPN as suggested, but the
login still timed out after ~1 minute with "Connection error." Right after
disconnecting the VPN, the app loaded fast again on my normal connection.
So it seems like the VPN test wasn't a clean comparison since the VPN's own
servers may have added latency. The core issue is still intermittent on my
direct mobile connection (without VPN) — sometimes loads instantly, sometimes
times out completely (ERR_CONNECTION_TIMED_OUT) for 60+ seconds, even though:
- DNS/CNAME is verified correct
-
- Container logs show healthy 200 OK responses during the same timeout window
-
- This also briefly happened on the auto-generated *.up.railway.app domain,
- not just my custom domain
Is there anything on Railway's edge/proxy side I should check, or could this
be specific to my region (India) reaching the US-East datacenter? Happy to
run any other tests that would help narrow it down.
0x5b62656e5d
I'm able to access your site. Try using a VPN. 
4 days ago
Thanks for checking! I tried connecting via Turbo VPN as suggested, but the
login still timed out after ~1 minute with "Connection error." Right after
disconnecting the VPN, the app loaded fast again on my normal connection.
So it seems like the VPN test wasn't a clean comparison since the VPN's own
servers may have added latency. The core issue is still intermittent on my
direct mobile connection (without VPN) — sometimes loads instantly, sometimes
times out completely (ERR_CONNECTION_TIMED_OUT) for 60+ seconds, even though:
-
DNS/CNAME is verified correct
-
Container logs show healthy 200 OK responses during the same timeout window
-
This also briefly happened on the auto-generated *.up.railway.app domain,
not just my custom domain
Is there anything on Railway's edge/proxy side I should check, or could this
be specific to my region (India) reaching the US-East datacenter? Happy to
run any other tests that would help narrow it down.
4 days ago
Try enabling CDN caching. Deploying to multiple regions also helps.
0x5b62656e5d
You can try deploying to other regions (eg, EU or SG).
4 days ago
Thanks for the suggestions! I'm planning to migrate both my API service
and PostgreSQL database together to Singapore region to keep them co-located.
Quick question before I start: is there a way to change region on my
existing services in-place, or do I need to create new services in
Singapore and migrate data over (pg_dump/restore) then cut over DNS?
Want to make sure I'm not missing a simpler official migration path.
ionixautomation
Thanks for the suggestions! I'm planning to migrate both my API service and PostgreSQL database together to Singapore region to keep them co-located. Quick question before I start: is there a way to change region on my existing services in-place, or do I need to create new services in Singapore and migrate data over (pg_dump/restore) then cut over DNS? Want to make sure I'm not missing a simpler official migration path.
4 days ago
Yes, you can change the region without creating a new service.
Open your project, select the service, then go to Settings → Scale and change the deployment region to Singapore. Railway will redeploy the service in the selected region.
You can also deploy the service to multiple regions at the same time if you want to improve latency and availability for users in different locations.
Just make sure your PostgreSQL database is located appropriately as well, since keeping the API and database in the same region generally provides the best performance.
sumitdynamic99
Yes, you can change the region without creating a new service. Open your project, select the service, then go to **Settings → Scale** and change the deployment region to Singapore. Railway will redeploy the service in the selected region. You can also deploy the service to multiple regions at the same time if you want to improve latency and availability for users in different locations. Just make sure your PostgreSQL database is located appropriately as well, since keeping the API and database in the same region generally provides the best performance.
4 days ago
Thank you, I'll do it.
4 days ago
Thanks Everyone! Does changing region under Settings → Scale migrate the existing
data automatically, or does it create a fresh empty instance in the new
region? I have production data I can't afford to lose.
0x5b62656e5d
It automatically copies the volume over.
4 days ago
Thank you