CONNECT_TIMEOUT Issue
Hello all, had an issue with a sporadic CONNECT_TIMEOUT across many services connecting to the same DB for about a 5min span a few nights ago. This seems to have happened again for less than a minute last night. Wanted to see if there's any way to debug this kind of issue?
My DB doesn't auto-sleep, and there's no indication anywhere on my dashboard that there was an outage. Oh, also I'm using the provided pooler/pgbouncer URL.
I have multiple services running in various locations that all had their DB connections timeout at the same time-- this is what gives me confidence that the issue was with Neon.
4 Replies
ratty-blushOP•2y ago
Reposting from https://discord.com/channels/1176467419317940276/1176467419938701375/1195031800188239902 in general since nobody from the Neon team responded 😅
Happy to share my project ID and stuff if useful
Main concern I'm having is that I'm building a service I want to have a high SLA, so I want to understand
1. was it really neon that failed? (with special interest in being notified when it happens again)
2. if so, why did it fail? (with special interest in fixing it so it doesn't)
3. how can I tell? There's no logging or anything on the dashboard that reflects "your DB went down for X seconds for Y reason"
I'm hoping these have reasonable answers rather than lead me down mystery rabbit holes where I have to debug my infra.
dependent-tan•2y ago
Hey @sina , sorry to hear about the connectivity issue.
Better observability features in our Console are on the roadmap, and should help with these sort of questions.
This question is best sent to support@neon.tech. They will be able to pull the logs and find specific information for you.
ratty-blushOP•2y ago
Ack, will repost this in the email then 🫡 thanks for the quick reply
dependent-tan•2y ago
You're welcome!