I have the same issue. I set up a managed Postgres cluster and I want it shared by a couple of App Platform apps, so I want to restrict access to just those apps.

After you’ve deployed your App Platform app, I found that you can go to the Console tab and run ifconfig and look for the IP address listed under inet addr.

Doesn’t seem very useful if we can’t guarantee that it will always have the same IP, though.

If I could create a managed Postgres instance in the same VPC as my App Platform applications then I could use the private connection hostname instead of the public one, but it doesn’t look like we can group App Platform and other managed resources into the same VPC, unless anyone knows different?

You might be able to find the public IP using a tool like OpenDNS’s myip by running dig +short myip.opendns.com @resolver1.opendns.com.

That said, I would caution you in using this approach for building firewall rules. First, and most importantly, traffic from other App Platform apps may egress with the same public IP address. Also, there’s not necessarily a 1-to-1 mapping of app instances to public IP addresses. App Platform may reschedule containers if the underlying host becomes unhealthy, for maintenance, or rebalancing resources. When possible, to avoid downtime we first start the new replacement containers before terminating the old. This means two or more containers may be running for a brief period. We also use a similar strategy when rolling out new deployments of your app.

For this same reason, we cannot assign a floating-IP or other stable IP directly to your application. It defeats the technologies we use to keep your application up and safe.

Our long-term plans include VPC support for App Platform which would let you communicate with your database on a private network. We’re also considering allowing App Platform apps to be specified as trusted-sources in the managed database product.