what considerations to use when building a microservice springboot app that will run in about 100 pods and they all need high speed SQL (read/write) access to an external PGsql cluster? will having each pod/container have its own jdbc connection pool which connects direct (as a service) to external pgsql cluster be wise to do? or is there a a better approach with a proxy-data layer or something of sorts?
Doubt anyone is going to have a real answer to this for you. Best advise I can give you is perform testing, make synthetic workloads happen against the cluster, and just see what happens. If there are bottlenecks, they will become apparent under load/stress.