From where are the metrics from which the ATC Experience Card Histogram Bar Chart bars drawn?

book

Article ID: 225116

calendar_today

Updated On:

Products

CA Application Performance Management SaaS DX Application Performance Management DX APM SaaS CA Application Performance Management (APM / Wily / Introscope)

Issue/Introduction

How are the values for the counters in the histograms team center updated ?

1) can alerts be raised when count represented by any individual 'bar' crosses a threshold?

2) can the number and definitions of bars be varied - so for example have a bar for experiences longer than 20sec?

Environment

Release : 20.x, SaaS

Component : APM

Resolution

The counters are incremented based on the responses per interval and the average response time for that interval - this is beacause the system does not have all the individual transactions response times so can only work on the aggregated values for each 15s interval

Number of transaction in 15 second interval (Responses Per Interval metric) is put to one of these buckets:

  • when Differential Analysis alert is up or an alert of Average Response Time metric is up, the RPI is added to "SLOW" bucket
  • when ART < 1 second, it is added to "1s" bucket. When an alert is up, it is shown in the bucket as red
  • when ART is between 1 and 2 seconds, it is added to "2s" bucket. When an alert is up, it is shown in the bucket as red
  • when ART > 2 seconds , it is added to "2s+" bucket. When an alert is up, it is shown in the bucket as red


metric Errors per Interval is added to "ERROR" bucket and Metric Stalls is added to "SLOW" bucket. It means "SLOW" buckets contains all stalls + RPIs when certain alert are up

 

The above behaviour is currently fixed and alerting cannot be done on the values in those histograms , just on the metrics that feed them.

Attachments