Skip to content

Latest commit

 

History

History
101 lines (68 loc) · 2.41 KB

customize-pcf-metrics-2-0.md

File metadata and controls

101 lines (68 loc) · 2.41 KB
Raw

How to configure custom metrics in PCF App Metrics 2.0?

Refer: Steps to register metrics endpoint

  • Install plugin
cf install-plugin -r CF-Community "metric-registrar"
  • Register a Public Metrics Endpoint

Note: that the metrics endpoint gets bound to app as service instance

cf register-metrics-endpoint adperfsimulator /actuator/prometheus
  • Register a Structured Log Format
cf register-log-format adperfsimulator json
  • Check registered endpoints
cf registered-metrics-endpoints

PCF App Metrics 2.0.x should have been installed, which is dependent on Metrics Store.

---
apiVersion: indicatorprotocol.io/v1
kind: IndicatorDocument

metadata:
  labels:
    deployment: "app_metrics"

spec:
  product:
    name: org,space,app-name
    version: 0.0.1

  indicators:
    - name: "JVM_Memory_PSOldGen_Indicator"
      promql: "max(max_over_time(jvm_memory_used_bytes{ source_id=\"$sourceId\" }[60s]))"
      documentation:
        title: "JVM - Memory Used in bytes"
      presentation:
        units: "bytes"

  • value for units: Should be one of none, bytes, counts, percentage, makeItAPercentage

  • Curl the doc

curl -vvv https://appmetrics.sys.domain/indicator-documents -k -H "Authorization: $(cf oauth-token)" --data-binary "@pcf-metrics-indicator-doc.yml"

# expect HTTP status code 200

Launch PCF App Metrics 2.0 and view dashboard.

PCF App Metrics 2.0 - Custom Dashboard

  • To understand the metrics and customize Promql

View actuator end point https://adperfsimulator.domain/actuator/prometheus

cf tail adperfsimulator -c metrics -f | grep jvm_gc

Note: for cf tail install plugin cf install-plugin -r CF-Community "log-cache"

How to configure Multiple App metrics in single dashboard?

  • Get guid of the secondary app which you want to include in your dashboard
cf app APP-NAME --guid
  • Replace $sourceId with guid of the secondary app
- name: "HTTP_Request_Count_UI"
  promql: "sum(rate(http_total{source_id=\"edcaf4c3-0a90-55dc-8726-00b86a5e226f\"}[367s]) * 60)"
  documentation:
    title: "HTTP Request Count (per minute) - Front"
  presentation:
    units: "counts"

Back to Table of Content