It has not been terrible but I won't run it all the time in production. It is better than doing (and reading) heap dumps for memory analysis. We recently did a benchmark/load test of our production environment and we used jmelody for analysis. It work rather well and we didn't see a significant "observer effect" impacting our max throughput. I don't have exact numbers. It would be interesting to seeing what the "exact" impact is.
tom On Sat, Aug 10, 2013 at 9:43 AM, Tom Poage <[email protected]> wrote: > Interesting idea, esp. references to e.g. rocksteady for large(r) scale > monitoring/analysis. > > Re: 'observer effect', any sense of the impact of jmelody on system > resources (server load, memory, ...)? > > Thanks. > Tom. > > On Aug 9, 2013, at 4:46 PM, Tom Freestone <[email protected]> wrote: > > We added jmelody to our CAS server to see ehcache stats. It is > straightforward to install. Here is a link > > https://code.google.com/p/javamelody/wiki/UserGuide > > > > tom > > > > > > On Fri, Aug 9, 2013 at 9:25 AM, Tom Poage <[email protected]> wrote: > > Anyone successfully get Ehcache to record/report cache hits/misses > (JMX)? Trying to see if my in-memory TGT cache is sized appropriately > (overflowed to disk). > > > -- > You are currently subscribed to [email protected] as: > [email protected] > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user > > -- ***************************** Tom Freestone ([email protected]) ***************************** -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-user
