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

This product reached the end of support date on March 31, 2021.

memory used by process doesn't match heap usage in the monitoring dashboard

Alejandro_soto
Inactive

Hi folks,

Is this normal?

I am seeing one of my processes (reports 2) using 470mb of memory as shown here:

but when I drill down to the process it shown committed memory of around 6Gbs.

Can someone please explain this?

I've checked other JVMs on different hosts and the values for process memory and committed heap match up for those... not for this one.

This is a JVM running in WAS on AIX.

DT version 6.1

Thanks,

7 REPLIES 7

Alejandro_soto
Inactive

After taking a memory snapshot, it looks like the number reported on the monitored processes is the used memory (470mb) as seen in the graph in dark red:

Why is there such a huge difference between commited and used memory and why doesn't that number agree with the used line in the heap dashlet (image2) which reports around 5.5Gb?

andreas_grabner
Dynatrace Guru
Dynatrace Guru

I would open a ticket to get the answer. could be a problem with capturing these metrics!

srikar_mohan333
Inactive

Would it be possible to create a single chart with these metrics (ensure all of them have the same unit MB) for the agent on jvm reports 2.

Under Server Side - Agent Based - JVM

used memory

committed memory

Under Server Side - Agent Based - Process Performance

current overall memory consumption

Per my understanding process memory reports on both heap and non-heap so it will be greater than heap only memory (jvm metrics).

Hi Srikar,

The opposite is actually true in my case:

I will open a ticket to see what's actually going on here and post back.

HI Alex, where you able to get an explanation/resolution for this? Thanks!

rvasuri
Newcomer

Hi - Is this issue resolved for AIS metrics next versions ? we are using 7.2 Appmon version but still see the issue


Ranganath, If you're still seeing this, I suspect it's an issue with how these metrics are reported by the container, likely something Dynatrace does not have control over the accuracy. To confirm this, I suggest opening a support case.