It should be possible to easily integrate Katja with existing pooling solutions like Poolboy or Pooler.
Since I want to keep Katja as lightweight as possible (dependency-wise), Katja will not take care of managing pools in any way (including pool membership). All Katja will do in regards to supporting pooling is:
Provide a way to not start/supervise the katja_metrics and katja_queries processes
Add APIs that do not depend on katja_metrics and katja_queries being registered processes
1. is basically already implemented in the development branch. The supervisor will look at the pool application setting and not add a process to its children if it's present in the poollist.
For 2. it will come down to adding APIs that take a PID (of a katja_metrics or katja_queries process) as their first argument.
It should be possible to easily integrate Katja with existing pooling solutions like Poolboy or Pooler.
Since I want to keep Katja as lightweight as possible (dependency-wise), Katja will not take care of managing pools in any way (including pool membership). All Katja will do in regards to supporting pooling is:
katja_metrics
andkatja_queries
processeskatja_metrics
andkatja_queries
being registered processes1.
is basically already implemented in the development branch. The supervisor will look at thepool
application setting and not add a process to its children if it's present in thepool
list.For
2.
it will come down to adding APIs that take a PID (of akatja_metrics
orkatja_queries
process) as their first argument.