> For the complete documentation index, see [llms.txt](https://hanfak.gitbook.io/workspace/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hanfak.gitbook.io/workspace/persistance1/connection-pooling.md).

# Connection pooling

* A pattern used with storing already precreated connections to a db, which can be used when talking to a database instead of connecting anew to a database.
* Similar to the pattern of object pooling

## Benefits

* speed, faster to use already created connections rather than alway creating new connections

## Libraries

* Java
  * HikariCP
  * C3po

## Improving

* More threads only perform better when blocking creates opportunities for executing.
  * As blocking because of waiting for IO to execute (thus more threads is better for old HDD rather than SSD)
* More threads than CPUs means context switching, due to time slicing, is slower than one thread per cpu
* You want a small pool, saturated with threads waiting for connections.
* <https://github.com/brettwooldridge/HikariCP/wiki/About-Pool-Sizing>