How Dynatrace calculate Linux memory usage

[ntanaka]

Hi,

I understand the Linux memory metric are read from /proc/meminfo.
https://community.dynatrace.com/spaces/482/dynatrac...

Now, I wonder what metric of /proc/meminfo dynatrace use to calculate Linux memory usage.

I guess the answer is following;

memory usage (%) = (MemTotal - (MemFree + Buffers + Cached)) / (MemTotal) * 100

Is it correct?

If it is different between Redhat Linux version 6 and 7, I would like to know the difference.

Best Regards,
Natsumi Tanaka

[ntanaka]

I am sorry to overlook this document
https://www.dynatrace.com/support/help/how-to-use-dynatrace/hosts/monitoring/measures-for-host-health/

It says;
memory_used = total_memory_size - (free_memory + active_memory + inactive_memory + reclaimamble_memory)

So, I guess these metrics can be replaced with /proc/meminfo metrics like following;
memory usage (%) = (MemTotal - (MemFree + Active+ Inactive + SReclaimable)) / (MemTotal) * 100

Do you think it is correct?

Best Regards,
Natsumi Tanaka

[ChadTurner · Accepted solution]

@Natsumi T. Dynatrace calculates memory usage as:

memory_used = total_memory_size - (free_memory + active_memory + inactive_memory + reclaimamble_memory)

we have seen this to be correct across all platforms

[ntanaka]

Thank you for your answer!