New postgres instance & listener issues

Hey gang - to take advantage of static outbound IPs, I had to recreate my infrastructure in a new Render team. This meant creating a new Postgres instance and restoring a backup from my previous Postgres instance, a version 13 Postgres instance I created about a year ago.

Ever since this migration, I’ve had persistent flakiness with Postgres LISTEN / NOTIFY functionality. I’m not aware of any major changes to this from v13 → v14. Were there any changes from Render around connection pooling or other middleware between my DB and my app that would cause these issues? Anything that would be part of a new DB created in the last two weeks vs one created a year ago?

For instance, pgbouncer is a common cause of LISTEN / NOTIFY problems. I don’t believe Render supports pgbouncer at the moment, but was there anything like this that was added in the last year that would cause these issues?

Hey Scott,
So creating a new team would actually mean that your services are now running in our AWS infrastructure as opposed to on GCP. There’s no PGBouncer or anything that’s been introduced.

Could you perhaps share some of the evidence of ‘flakiness’ - if you don’t want to share them here on a public forum please feel free to email them through to


John B

So creating a new team would actually mean that your services are now running in our AWS infrastructure as opposed to on GCP

Whoa. If there was an announcement of that, I definitely missed it. That explains some recent unrelated strangeness in my setup then - I have a decent amount of supporting infrastructure in GCE outside of Render, and it’s all performed miserably since I recreated my stack. You definitely saved me some time on unrelated troubleshooting :slight_smile:

I’ll let this thread close. The move to AWS means I need to start migrating off the Render platform unfortunately, so this is not worth further troubleshooting.

One last note for posterity - since I still had suspended, old GCP instances, I redeployed my code and let both run in parallel for a few days.

GCP & Postgres 13 - no issues
AWS & Postgres 14 - no new NOTIFY events after about 12-24 hours following a new Postgres connection. (e.g. I can restart my app and it works at first, then dies)

I wish there was more detail I could provide, but Postgres doesn’t provide a lot of troubleshooting information to run this down, and from my application’s perspective all is well – it just stops getting any new notifications from Postgres.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.