Supabase Instance Timing Out For Two Days, no changes made — Possible IOPS or Pooler Issue?
Starting Thursday around 6 PM UTC, my Supabase instance began timing out frequently. My data ingestion pipeline has been stable for several weeks, and I’m unable to determine whether the issue originates from my application or from the Supabase infrastructure.
Queries that do go through are not slow (most are under 200ms), and the dashboard shows no high disk I/O or CPU usage. Tables are well indexed and queries are optimized to be as light as possible. I’ve tried setting idle_in_transaction_session_timeout, doubling compute size, vacuuming large tables, and reducing query load, but the problem continues. It seems like something is stalling the pooler or connections, but I can’t find the root cause. I've attached a Grafana screenshot, you can see the connection drop out regularly and ping until reconnecting. I can share logs if helpful. Any ideas to troubleshoot or is this a Supabase issue? I have contacted support but haven't heard back. Thanks in advance!
3 Replies
Does not seem to be something anyone else is hitting suddenly. So either your queries, overloading some resource or some issue with your instance.
thanks! Im not seeing any queries that are slow in query performance and have no indication on my supabase stats that anything is getting overloaded
There was an issue in US-west over the weekend but you were before others were having issues.
Restart your instance?
If you use PostgREST API check API report to see if any unexpected high volume queries.
Check general logs for timeout errors.
Otherwise you'll have to see if another user has ideas for you.