When should multiple Cloud SQL proxies be used to a single instance of Cloud SQL for Postgres?
We currently are considering doubling Postgres max connections from ~2500 to ~5000 due to running out of connections and would like to know what are the considerations for adding another Cloud SQL proxy.
The number of queries hitting our Cloud SQL Postgres is high - it can be well over 500K queries per minute. The Cloud SQL Postgres instance has 40 CPU, 250 GB of RAM. It appears that not all of the memory is being used - approx. 90 GB is still available. Any advise would be appreciated.
The proxy does not place any limitations on new connection rates or persistent connection count. However, you can connect the proxy to multiple instances, feel free to refer to the docs.