Our application on Cloud Foundry creates many connection pools to the same Postgresql database, each pool is to a different schema.
I found that every time the application crashes when it's about to create pool number 350 something. (e.g the 356th pool) The crash is due to error code 137, which means it's likely an OOM issue. This is happening while there are plenty of memory left unused.
Tried a few configuration combinations with the memory calculator,
such as increasing the stack_threads and -Xss but to no avail.
Searched around but couldn't find anybody talking about seeing the same problem, so is there a maximum number of connection pools that HikariCP supports? If I switch back to Tomcat Connection Pool there is no such problem.