T
tbh
cpu load increased on our web servers unexpectedly after a recent, fairly
routine "release change" (which is to say, we changed a bunch of our scripts
and libraries in one fell swoop). i can back off to the previous release,
but it is typically quite hard to simulate live load on test systems, so in
general it would be good to know in a live setting where this additional cpu
drain is coming from.
i know which application is using the most cpu, but that was expected and
doesn't help me much.
is it possible to measure how much cpu one request used and record it to a
log when it ends (in Application_EndRequest())?
i looked quite a bit just now and found papers on instrumentation and
profiling, but wasn't able to find anything this specific yet.
i will continue to look, but maybe someone can point me quickly to what i
need. i'd be grateful.
cheers,
Tim Hanson
routine "release change" (which is to say, we changed a bunch of our scripts
and libraries in one fell swoop). i can back off to the previous release,
but it is typically quite hard to simulate live load on test systems, so in
general it would be good to know in a live setting where this additional cpu
drain is coming from.
i know which application is using the most cpu, but that was expected and
doesn't help me much.
is it possible to measure how much cpu one request used and record it to a
log when it ends (in Application_EndRequest())?
i looked quite a bit just now and found papers on instrumentation and
profiling, but wasn't able to find anything this specific yet.
i will continue to look, but maybe someone can point me quickly to what i
need. i'd be grateful.
cheers,
Tim Hanson