KB 214 (ABAP): How to extract and send KPIs from collected logs

Description

In some case it might be needed to extract certain KPIs from processed data and send it as a metric to platforms which support metric ingest. Starting from PowerConnect SP 8.03 it is possible to use Administrator->Setup Metric->KPI Definitions to achieve that.

in SP 8.03 it is possible to send KPIs as metrics to following endpoints:

1. Splunk Observability (sent as OpenTelemetric datapoint)

2. Splunk HEC (send as regular event)

3. Dynatrace native metric endpoint

4. Dynatrace OpenTelemetry endpoint (sent as OpenTelemetric datapoint)

Resolution

Following steps are required to create the KPI definition:

Create dedicated KPI subgroup (if necessary)

1. Open Administrator->Setup Metric->KPI Definitions
   

   

2. Double-click on Custom group in the left panel and click.
   

   

   This group should be always used for custom KPI definitions.

3. Populate group details based on the context, set Active flag and confirm

   

Create KPI definition

1. Select the group by double-clicking on it. The title should be changed to the name of defined group
   

   

2. Click on Create KPI button Populate
   Name: should be in lowercase and not contain spaces or any non-readable characters
   Description: free text to describe the KPI
   Active: used to activate/deactivate KPI extraction
   Applicable for section is used to select Group Definition/event type and Subgroup/subtype, which will be the source of the KPI data
   Calculation Rule section contains the rule on how to create KPI value.
   
   Field: the field name, which exists in the selected Group/Subgroup defined in the section above
   Dimensions: used to aggregate data per certain dimensions like Instance name. Indicated field should exists in the selected Group/Subgroup defined in the section above
   
   Function: aggregation rule used to generate the value. Following values/functions are supported:
   None - no aggregation made and the value is taken as is from each event (can generate a lot of metrics in case there is big dataset)|
   Average - average value is calculated
   Sum - total sum
   Max - maximum value
   Min - minimum value
   Count - count of events in the batch
   
   Unit: free text, which would describe the unit
   

   

Product version