Do we have memory analysis capability in dynatrace saas or or Managed like we have in AppMon.
one of my process old generation getting filled up very quickly, gc triggering and it is taking very long time due to which process getting suspended.
In the problem pane it show CPU saturation and root cause is Long garbage collection suspending process.
Can we do in depth memory analysis in dynatrace.
Solved! Go to Solution.
As Michael teased above, this feature is in the works and will be arriving shortly. As a suggestion, subscribe for updates to the blog and product news at the bottom of this page: https://www.dynatrace.com/blog/
Enter your email address and you will be notified as soon as this feature is available. This also gives you an area to view the feature and read all about what you can do with it.
Its out https://www.dynatrace.com/blog/analyzing-memory-dumps-for-java-and-node-js/
Thanks I got it
Related question: During our long memory leak load testing phase we utilize the trending memory dumps with GC enabled (similar to jmap -histo:live). The current appMon trending memory dumps features to compare these "jmap" listings and filter is a feature we utilize for every app release. Are those features available in the managed dT?
not yet, we do have further memory analysis capabilities on our roadmap.
Thank you Michael.
One approach/thought I had was ... would there be value in packaging the dT AppMon trending memory dump analysis tool and contribute to the community as an independent tool? This would make the key trending data analysis/filtering/sorting features available to all. Maybe this then simplifies your dT managed feature to simply be to schedule/collect jmap output files.
I don’t current work with IBM or other JVMs to know if jmap –histo:live works now/in-future for non-Oracle/OpenJDK JVMs so this might be a serious flaw in my thinking.
Otherwise then in dT managed you could reduce the existing trending feature to have dT Managed remotely issue the jmap -histo:live (or - :live), +collect, +ability to download local … somewhat similar to the current dT Managed approach (my limited understanding =>) to just schedule/utilize normal heap dumps +/-, providing a way to collect/download them, and then just use existing community dump analysis tools such as MAT/visualVM with the hprof/phd heap dump files.
I do believe your current dT AppMon trending heap dumps analysis features would be very useful as a java community contribution if it could be broken out and packaged independently. Maybe included in MAT or as a separate eclipse based package.
I hope my description of my idea is worded well and is useful in some way.
I had this package independently idea a long time ago and often since at our location there are critical places where dT AppMon is not permitted at all (crazy … I TOTALLY agree, /cry /cry /cry) and so having the ability to do analysis across several collected jmap –histo files would be very handy to have when not having dT AppMon agents in place. Instead I have to let the app leak for some while and then attempt to get prod team to take a full heap dump some night and then hike Everest 10x to deal with sftp accesses on some (rarely) commonly accessible host in order to get the full dumps to me (easier to send to vendor = v weird/crazy too) for loading in MAT. The jmap histo files could be emailed to me and get them sooner. But without your trending analysis tool features I just resort to wait for the full heap dump (which does have best data when I can actually get them). So … it is a weird combo of stupid crap at my company that caused me to want this as an independent community feature and therefore never asked … but had I known the future I would have surfaced this idea long ago.
Thanks if you read this far ... have a great day!
Is this feature out yet? if yes, which version of Appmon has it? 7.0/7.1?