I might be able to comprehend how the network metrics are being calculated for the NICs of host level, but how about process level?

[waikeat_chan]

How did the calculation of network metrics on process level is being carry out?

FOR example, this screenshot:

How can the AI tell? And how can it tell it isn't the host level metrics that's impacted but just the process level?

[pawel_brzoska · Accepted solution]

Hi there,

one agent is able to understand, dissect and associate network conversations to individual processes. We do it by looking at individual network packets, not only at interface stats.This allows us to precisely pinpoint network issues and related them to individual monitored process. This, in turn, allows AI to bound this event with performance problems of the whole app that depend on the process having network communication issues.

[tibebe_m_digafe]

Hello,

Not exactly related but appreciate any insights as to what should I enable to see the network metrics for processes. A managed offline cluster in my case.  I only see the metrics for some processes such as filebeat, kubelet .. but none for other processes, the vast majority.

Dynatrace folks can look at case#17511

Thanks. Tibebe