ARQ supports JMX for a few things:
* Static information - version of software running
* Live information - count of queries and some details of the last query
run
so it isn't very much information and it is fixed. It was my plan to
expand this if there was demand. Logging is possibly a better way to
get details of the last query. (Try it with jconsole.)
But JMX is problematic in various environments (various reports in the
last 2 weeks: OSGi/glassfish ([*]), Google app engine (JENA-613),
running schemagen in ant (because of software version registration of
RIOT) [offlist]).
Thinking of Jena (core, ARQ, RIOT) as a library + framework then maybe
it's better to target the wider range of uses and regard these sort of
statistics as better done within the application using Jena because it
has an understanding of what functionality is being offered and why.
Fuseki keeps separate, and much more detailed, statistics based on
services, good and bad requests. These could be exposed as JMX, and be
better information than currently available but it's better done moving
the JMX code out of ARQ into Fuseki iff needed.
Does anyone use JMX with Jena? or know if it's use?
If it's unused, then it's more trouble than it's worth and I propose
removing it.
JENA-614
Andy
[*] http://markmail.org/message/he5koceg6spznlbi
