Currently, one sends mgr queries to the "whole" Squid. Kids responses may get
aggregated by Coordinator, and we, in general, want to aggregate all responses
that can be aggregated.
This patch allow us to aggregate stats for a subset of kids. For example,
the following query aggregates stats for just the first and the third workers:
mgr:info?workers=1,3
When query response information cannot be aggregated (or at least is not
aggregated right now), then a parameterized query will result in several
matching "byKid { ..." blocks.
This patch support the following scope variants:
* raw interface with access to any kid process or groups of kids; similar
to ${process_number} macro we already support in squid.conf:
mgr:foo?processes=id,id,id...
* higher-level interface to isolate workers by their numbers, starting
with 1 for the first worker:
mgr:foo?workers=num,num,num...
Currently, all kids except Coordinator are workers, but that will change in
the future as we get more kinds of kids.Currently, one sends mgr queries
to the "whole" Squid. Kids responses may get aggregated by Coordinator,
and we, in general, want to aggregate all responses that can be aggregated.
This is a Measurement Factory project