Hi, We are currently monitoring production application with load around (throughput 1,20,000 per min).The dynatrace server sizing is xLarge96. We are facing problem where throughput suddenly goes down to zero and dynatrace server stops analysis of purepaths resulting in no data in charts. At the same time purepath buffer memory gets full and lag duration starts increasing.
We also observed some purepaths with high size upto 7000 to 8000 nodes during.What could be the possible cause?
Thanks in advance.
We had issued the same situation in the past. I used to check the "server health" dashboard in order to know how are those purepaths beeing lost...
Thousands of nodes per purepaths is of course an issue. If 10 purepaths with 10000 nodes arrive, the limit is over and you start to not collect those executions.
In any case, try to see if the purepaths are beeing early discarded or whatever, in order to see where the bottleneck is (server, colector..)
Hope this helps!
Thanks Arturo for your reply. we did not observe skipping or early discarding of purepaths during such incident. Dynatrace server just stops showing purepaths. Also to add, Can we decrease maxnode limit from default value 10,000? If not, what could be the alternative.?
I would see those large executions and which sensor/s are responsible for that amount of nodes...Usually DB/cache related sensors do this kind of things, but maybe others such hibernate could be the ones on your case!
Regarding the bottleneck on your system, you should check carefully the pieces involved. Maybe your collector can't manage that amount, or the server itself. Just try to go further and try to see where the problem is!
In any case, reducing the amount of nodes per purepath will help for sure!