Long story short, the memory leak is so serious that I even try to configure actions which are triggered full dump once it Memory Usage hit 80% (my first attempt), follow by 70% (second attempt), then 50% (the third one).
But even with configured to triggered the full dump during 50%, still didn't succecced. To complicate things, the period to let the problem be reproduced/appear again is somehow.....not periodic.
Which means, for the past month there already are 3 incidents (and hence, 3 restart of the JVM) and yet I still don't have a full dump. Memory leak is so serious (and happens so fast) that during the memory usage spike, I can't even trigger a trending dump, agent simply dies once I triggered trending dump, only to wait for it to come back alive after a while.
So while waiting for the fourth incident (and a lower threshold to triggered a full memory dump) and cross my finger in hope of getting a dump successfully, I am wondering what else can I do, since God knows how long I have to wait, maybe tomorrow, maybe more than a week later.
Which brings us to this question, if this is possible I can ask the dev team to let me have their dump file, so that I can have a look in AppMon.
Unfortunately the format of the snapshot/dump used in AppMon is proprietary so dumps from other tools and such won't work as far as being analyzed in AppMon. You seem to have tried most of what I would recommend to try and get a snapshot, if it is this severe the analysis may need to be done outside of AppMon.
Here's an old RFE/Idea you can add to - won't help much immediately though.