22 days ago
We have multiple services in the same Railway project (vivacious-determination) that cannot make any outbound HTTPS connections. All affected services throw socket.gaierror [Errno -2] Name or service not known when attempting to reach external hosts.
Affected services:
wso-systems (always-on portal)
wso-healthcheck (always-on)
wso-intelligence (cron)
wso-email (cron)
All four services fail on outbound HTTPS. All four connect to Supabase (ysmxqkqzcmbtoihakua.supabase.co) as their primary external dependency. Supabase confirms their project is healthy.
Unaffected services in the same project making outbound connections successfully:
wso-ga4-pull
wso-meta-pull
wso-activecampaign-pull
wso-optimonk-pull
wso-ticketing-pull
wso-sync-ticketing
All services are in US East (Virginia). Variables are confirmed correct. The pattern appears to be: services that connect to Supabase cannot make outbound HTTPS. Services that connect to other external APIs work fine.
Outbound IPv6 is currently enabled. We can toggle this if it assists diagnosis.
Please advise. This has been down since yesterday.
4 Replies
Status changed to Open Railway • 22 days ago
22 days ago
socket.gaierror [Errno -2] is a DNS resolution failure. Try to disable outbound IPv6. If it's not the issue, we rule it out at least.
22 days ago
Disabling Outbound IPv6 made no difference. Same socket.gaierror [Errno -2] on all affected services. IPv6 is now ruled out. Any other suggestions?
22 days ago
Ok, to check whether the problem stems from your code or your container; I suggest you SSH into your service and run a command like nslookup or curl on your external service. If DNS is resolved correctly, then it's a code problem. Sharing the relevant code snippet can be helpful in this case.
22 days ago
Hi bro,
Please compare one affected service against one working service and verify:
echo "$SUPABASE_URL"
cat /etc/resolv.conf
getent hosts ysmxqkqzcmbtoihakua.supabase.co
curl -v https://ysmxqkqzcmbtoihakua.supabase.co
If DNS fails only in the affected services, this points to a Railway/container DNS issue for those services. But, if DNS works from the shell but fails in the app, then it’s most likely app config/code: malformed Supabase URL, extra quotes/spaces/newline in env vars, or the code passing the wrong value as the hostname.