Showing results for 
Show  only  | Search instead for 
Did you mean: 

How can I schedule a conditional memory dump using task?

Using task, I can schedule AppMon to do a memory dump at let's say 8pm everyday.

On the other hand, using incident rule, I can make a memory snapshot to be done let's say, when the memory usable has become less than 35%.

However, if I want to do a memory dump only if when both conditions .i.e

-when the memory usable drops to below 35%

-the time the incident happens is after office hour (for example, when the time is in between 7pm to 9pm)

Is it possible to do so? (means, don't want to dump if its still office hour, although the memory has violate threshold)



Hi @Wai Keat C.

Using Incidents we can take the Memory dump, if the particular measure(only agent based measure such as.NET Memory consumption(Heap Size Gen1), Memory Utilization(JVM)), etc..) exceeds the defined threshold, for this we use Memory Dump Action Plugin to achieve this.

Memory Dump Action Plugin gets executed on the particular agent side if exceeds the thrshold of the measure which we configured in incident and stored in Session storage.In Cockpit, if you click on Session storage, you can see Memory dump created by the incident at the specific time(when memory exceeds the threshold).

----Using task, I can schedule AppMon to do a memory dump at let's say 8pm everyday----

Main purpose of memory dump is to analyze the memory usage if any memory leak/memory issue happening which indirectly affects the methods/process running in the application servers, which in turn affect Application performance/usage.

Using task, I don' t believe there is a way to schedule memory dump at a particular time because no purpose for such a feature needed in dynatrace-appmon i guess.

Hope this may help!!


Soorya Mohan

Dynatrace Leader
Dynatrace Leader

I think you should be able to get this behavior. Once you have your incident created you have the option of setting a "downtime" for the incident in the incidents dashlet. So if you set there to be a recurring downtime during your office hours the incident and thus the memory action will not occur during those times.

Since you'll still likely want to be alerted when the memory incident is violated I would actually create a second incident that is identical to the original (you can just copy-paste) with the only difference being the second is the one with the memory dump action and the office hours downtime.

Does that help?



I would try that later, but that sounds to be a simple and elegant solution.

Thanks for the suggestion 🙂

Wai Keat