So I’ve been using this workflow for a couple of months but approximately a week ago, the PostgreSQL node stopped working. When I wanted to troubleshoot it, it says this error message that says “connection timeout”.
I use an N8n docker hosted into a Render server and I use a Supabase database with a basic table that saves àmy session IDs and the output and inputs for the AI to remember the conversation. Since it stopped working a week ago, I’ve tried to troubleshoot it for a couple of hours yesterday.
I updated my N8n version to 2.8.3; it didn’t work.
I updated my database on Supabase; it didn’t work.
I also tried to change my port for my PostgreSQL for session pooler 5432 instead of transaction pooler 6543 , but I have the same error.
Even in the successful workflows, they successfully retried a couple of times but they didn’t get access to the memory. The PostgreSQL doesn’t work so I don’t know what to do with that because every other connection through API, even my super base ones that I connect through OAuth, works. It’s just the PostgreSQL that doesn’t work and I don’t know what to do with that to make my workflow work.
Also I want to mention that when I delete and recreate my credentials sometimes it loads infinitely and it doesn’t connect. I need to refresh my page and try to re-do the connection to my Postgres credentials. It’s not just an error of workflow; it’s an error with the credential and the connection.
the username is usually what breaks it. it cannot just be postgres. it has to be postgres.[your-project-ref]. copy it exactly from the supabase dashboard.
Hi @Samuel_Ross
Welcome to the n8n community
Supabase automatically pauses free tier databases after 7 days of inactivity. The timing you mentioned (“about a week ago”) matches this exactly. Head over to your Supabase dashboard and check if your database shows a “Paused” status. If it does, just click to restore it and try your connection again. That might be all it takes!
Ah got it, that rules that one out!
did you get a chance to try what @A_A4 suggested? If you already tried that and it still didn’t work, a couple other things worth checking:
First, on the Supabase side, do you have any Network Restrictions enabled? If Render’s IP changed recently (which can happen), Supabase might be blocking the connection silently, which would explain the timeout instead of an auth error.
Second, the fact that it loads infinitely when you try to recreate the credentials is a red flag for a network/firewall issue rather than a wrong credential. A wrong password would just fail instantly.
looking at your screenshots I think I spotted two things.
First, the Supabase pooler actually requires SSL to work, but the trick is to enable SSL and also turn on “Ignore SSL Issues (Insecure)” at the same time. Just enabling SSL alone without that option tends to break things in hosted environments like Render.
Second, 15 max connections is probably too much for Supabase’s free plan. Try dropping that to 3 or 5 and see if the intermittent failures get better.
