What is the use of count in stackdriver metrics

Refresh

March 2019

Views

230 time

1

I have a log metric that is like this:

description: hangup
filter: |-
  resource.type="global"
  jsonPayload.message="hanged up"
labelExtractors:
  campaign: EXTRACT(jsonPayload.context.campaignId)
  causa: EXTRACT(jsonPayload.context.attempt.hangupCause)
  pickedUp: EXTRACT(jsonPayload.context.pickedUp)
metricDescriptor:
  description: hangup
  labels:
  - description: causa de colgado
    key: causa
    valueType: INT64
  - description: "campa\xF1a"
    key: campaign
    valueType: INT64
  - key: pickedUp
    valueType: BOOL
  metricKind: DELTA
  name: projects/redacted/metricDescriptors/logging.googleapis.com/user/hangupcause
  type: logging.googleapis.com/user/hangupcause
  unit: '1'
  valueType: INT64
name: hangupcause

If I do an aggregation/alignment by SUM it does what I expect in an aggregation by count:

SUM

IF I do COUNT for example in ALIGNMENT/AGGREGATION :

COUNT

I want to understand what is the use of the COUNT function in alignment and aggregation? What are we counting or summing ?

1 answers

2

In GCP, “count in aggregation” is the active number of metrics running at that particular time. For example, in your diagram, the red line is describing that for that particular group you are monitoring, there are four active metrics running.