how to get normalized builtin:host.cpu.load5m

[Vijayt]

i have access to only classic dashboard . 

builtin:host.cpu.load5m  - shows overall . server load 

i am want to create alert if builtin:host.cpu.load5m/ cpu count  >1.5  but i dont see any option to use cpu count in metric. logical cpu cores is in properties. how to use that in data explorer

[AntonPineiro]

Hi,

If metric is not provided by OneAgent, you have different ways to ingest metrics in Dynatrace.

Scripting integration is easy since you have already OneAgent installed.

Best regards

[Vijayt]

Cpu count is static and our query based on time based.  do i need use dynatrace_ingest to send cpu count. every minute?

[AntonPineiro]

Hi,

It depends what you want to do. If you only need alert, and no metric. You can ingest an event when CPU is higher.

Or if you are in Grail, using DQL + workflow for checking it, one time per day, and generating a problem.

Best regards

[t_pawlak · Accepted solution]

Hi,
You can’t do builtin:host.cpu.load5m / cpu_cores in classic dashboards/Data Explorer, because logical CPU cores is a host property, not a metric — and metric math there works only with metrics.
IMO You should do what @AntonPineiro suggest. Iingest CPU cores as a custom metric (e.g. custom:host.cpu.logical_cores) with dt.entity.host, then compute load5m / logical_cores. Send it every 5–15 minutes (same value) so alert evaluation doesn’t hit gaps.

Alternative: use built-in host anomaly/CPU overload detection instead of normalizing load.

AIX edge case: AIX may have built-in “core/capacity” metrics, so you can normalize without custom ingest.

Best regards

[t_pawlak]

If your goal is just alerting, check host anomaly settings already include a threshold.
https://docs.dynatrace.com/docs/observe/infrastructure-observability/hosts/configuration/anomaly-det...

[Vijayt]

Thank you for detailed explanation but normalization will  be useful if taking historic trend