Well I don't know what was the plan at first, but in the code, it was used for the maximum number of connections per host.
If it was for the maximum number of WS clients for a given operation, then it would be settable per Provides and not per Component.
For now I didn't change the current behaviour, that is to have the same max number of ws client per service than the max number of processors.
In some case, we could find ourselves with more ws clients that usable (since there will be many services per components), for example just after a lot of call to the same service happened. But I think it's an ok default situation and after some time the extra ws clients will be purged from the pool and things will be ok
We could think about introducing a parameter to set the ws client pool (actually named ws-clients-pool-size-max but PER PROVIDES and not component), but I'm not really convinced it is so useful…
Are you sure that ws-clients-pool-size-max is the maximum number of connection per host ? IMO, its the maximum number of WS clients for a given operation of a web service, that is not the same thing.