Release : 22.2.x
Component : IM Polling
When any given metric family takes over 50% of a poll cycle to complete, the threshold monitoring engine transitions to a degraded state.
At this point an event is generated on the Data Aggregator indicating that if the evaluations continue to take over 50% of a poll cycle,
threshold evaluation will be disabled if this persists for 15 minutes.
Create an on-demand report for the Data Aggregator component for the following 2 variables.
We want to show the metrics by component so set the view type to be "Chart per Item with Multiple Metrics" then set the "Metric Calculate Level" to By component.
Data Aggregator Event Calculation Times : Percent Of Poll Cycle To Complete - Maximum
Data Aggregator Event Calculation Times : Number Of Processed Timestamps - Average
Run the report for around the time frame that the the threshold sate was degraded
The report will show you which metric family is causing the event processing to take a long time.
You can then adjust the amount of event thresholding you are doing on that metric family to reduce the load.