topic Aggregation on metrics exposed through API in Dynatrace API

Aggregation on metrics exposed through API

suresh230591 — Tue, 13 Feb 2018 11:19:49 GMT

Hi,

What is aggregation being used for problem pushed through API for infra, service, application when we are using automatic anomaly detection or Custom threshold for these metrics like Disk, CPU , Memory etc. Whatever exposed through API.

If I want to change aggregation for custom threshold , how can I do that

Regards
Suresh

Re: Aggregation on metrics exposed through API

wolfgang_beer — Tue, 13 Feb 2018 12:18:28 GMT

For slowdowns in automatic mode we alert on 50th and 90th percentile. Same for service response times as well as applications.

There is no way of changing the aggregation type for custom thresholds but what you can do is to define a custom alert where you have the option to choose between the aggregation types.

Re: Aggregation on metrics exposed through API

suresh230591 — Wed, 14 Feb 2018 04:51:45 GMT

Thanks Wolfgang,

Do we have any timeseries Id for Automatic Alerts generated by dynatrace.
for example CPU Saturation , it will create a problem if cpu usage cross 95%. But I did not find any timeseries for that. I can see timeseries for other cpu usage metric like stealtime, idle time, system usage but not the overall usage of CPU.

Regards
Suresh

Re: Aggregation on metrics exposed through API

wolfgang_beer — Wed, 14 Feb 2018 07:29:16 GMT

Disk aggregation that is used is the average. Availability for process is based on a process instance count where we use the min aggregation for alerting on the minimum instances running thresholds. For resources most of the time its also the average aggregation type.

Re: Aggregation on metrics exposed through API

suresh230591 — Wed, 14 Feb 2018 14:02:52 GMT

Thanks Wolfgang,

Do we have any timeseries Id for Automatic Alerts generated by dynatrace.
for example CPU Saturation , it will create a problem if cpu usage cross 95%. But I did not find any timeseries for that. I can see timeseries for other CPU usage metric like steal time, idle time, system usage but not the overall usage of CPU.

Regards
Suresh