Aurora Serverless (postgres) v1 down to 0 not scaling back up properly

Update: Turns out it is the call to SSM (Parameter Store) to get DB info that is timing out, so this is not actually related to Aurora Serverless.

Our staging environment is setup to scale down to 0 when not in use. This has been working great for a few years now until recently.
I only noticed this last week, but now we're getting inconsistent timeouts when it scales back up. Since we're using postgres, and the minimum is 2 capacity units, it almost feels like one unit boots up, and the other doesn't. Not sure if that's true, or there's something else weird about the scaling?

Steps to reproduce:

• Load a page that requests something from the Aurora cluster that is scaled to 0
• Wait for 30-60 seconds for the cluster to start
• Reload the page

Result:

• When I reload the page that hits the DB, it sometimes works, and sometimes doesn't. When it doesn't work, it times out (exactly as it did the first time). There's no pattern, and I haven't yet found a way (even after a few minutes of trying) to make it consistently stay up.

This makes CI (especially migrations) and smoke testing on staging painful, where i have to re-run the action several times before it succeeds (it eventually randomly hits the DB while it says it's up). This is not an issue on our production instance which doesn't scale down to 0.

Both instance are using:

• Aurora Serverless v1
• Postgres 11.18

Staging Scaling config:

• Autoscaling timeout: 5 minutes
• Pause compute capacity after consecutive minutes of inactivity: 10 minutes
• Minimum Aurora capacity units: 2 capacity units
• Maximum Aurora capacity units: 2 capacity units
• Force scaling the capacity to the specified values when the timeout is reached: Not enabled