topic Re: "Automatically" - anomaly detection in Open Q&A

"Automatically" - anomaly detection

jordan_rose — Thu, 05 May 2022 07:53:50 GMT

How can I get more info on what automatically means in the anomaly detection settings. I cant seem to find any docs on it and get those questions all the time.

For example - Detect slow-running disks 'Automatically' vs the easy to read 'based on custom settings'.

What does "Automatically" mean in this case?

Is there doc for what "Automatically" means for all of our anomaly detection settings?

Thanks,

Jordan

Re: "Automatically" - anomaly detection

andre_vdveen — Sun, 26 Jun 2022 17:07:31 GMT

Hello @jordan_rose, have you looked at the documentation page covering How problems are detected and analyzed?
Your example of slow-running disks is defined by static thresholds for infrastructure, which also covers CPU, memory, network, Garbage Collection, and Java OOM to name a few.

Dynatrace also uses automated multi-dimensional baselining for application and services error rate, response time, and load.

HTH,

André

Re: "Automatically" - anomaly detection

techean — Mon, 27 Jun 2022 05:47:20 GMT

Auto base lining is responsible for auto setting threshold for alerts when automatic anomaly detection is enabled. This needs reference set of date collectively aggregated for week to auto define thresholds for sending needful auto alerts on threshold failures.

Below are few links that explains it

Anomaly detection powered by AI | Dynatrace

Adjust the sensitivity of anomaly detection | Dynatrace Docs

https://www.dynatrace.com/support/help/shortlink/automated-baselining