Event throughput is not satisfactory

Edit online

For a production size environment with three datarouting pods, the event throughput is currently limited to about 50 events per second.

If you do not use the event reduction dashboard to view the related statistics, you can disable the pipeline that routes events to that consumer. This reduces the load on the event processing infrastructure and increases the event throughput to some extent. Perform the following steps to disable the pipeline for the event reduction dashboard:
  1. Get the datarouting deployment name:
    deploy=$(oc get deploy | grep common-datarouting | cut -f1 -d" " )
  2. Run the following command to list the activated pipelines:
    oc set env deploy ${deploy} --list | grep PIPELINES
    The output for an unmodified cluster should look like this:
    PIPELINES=ea-collatedactions,event-reduction,incidents,rba-execution-records,rba-statistics
  3. Remove the event-reduction pipeline by running the following command:
    oc set env deploy ${deploy} PIPELINES=ea-collatedactions,incidents,rba-execution-records,rba-statistics
Optionally repeat step 2 to verify that the event-reduction pipeline was removed. The datarouting pods should restart automatically. Afterward the event reduction dashboard will not receive further updates. Existing data is still displayed.