Skip to main content
Version: 1.5.0

prometheus

This guide shows how to monitor Apache APISIX Ingress Controller using Prometheus and Grafana.

Enable Prometheus#

Use CRD file to enable Prometheus in global configurations. The definition file for custom resources is ApisixClusterConfig, so the configuration should be:

apiVersion: apisix.apache.org/v2
kind: ApisixClusterConfig
metadata:
  name: default
spec:
  monitoring:
    prometheus:
      enable: true

Configure Prometheus Server#

The Prometheus server address should be 127.0.0.1:9090 by default. You can set the target url for apisix-ingress-controller manually in prometheus.yml. For example:

...
scrape_configs:
  - job_name: "apisix-ingress-controller"
    static_configs:
    - targets: ["localhost:9092"]

According to the above example, the metrics are exposed in http://localhost:9092/metrics now.
Visit http://localhost:9090 in your browser, and select apisix-ingress-controller in targets or visit http://localhost:9092/metrics to see the data.

Grafana Dashboard#

Grafana dashboard shows the metrics exposed in Prometheus graphically.
Installing Grafana
Visit http://localhost:3000/ in your browser to access Grafana. The default username and password are admin. Then create a new dashboard for apisix-ingress-controller.


Follow the steps to apply the configuration of Grafana Dashboard for apisix-ingress-controller.

  • Add an empty panel, and import via panel json in apisix-ingress-controller/docs/assets/other/json/apisix-ingress-controller-grafana.json.
  • Select Prometheus database as the datasource. Set the URL according to your Prometheus server configuration.

Preview#

Available metrics#

  • Is leader A gauge type metric with value 0 or 1, indicates whether the role of controller instance is leader, for leader is 1 and candidate is 0.
    Labels:
    • controller_pod
    • controller_namespace
  • Status codes status codes of requests to APISIX.
    Labels:
    • controller_pod
    • controller_namespace
    • status_code: the HTTP status code returned by APISIX.
    • resource
  • Latency Request latencies with APISIX.
    Labels:
    • controller_pod
    • controller_namespace
  • Requests Number of requests to APISIX.
    Labels:
    • controller_pod
    • controller_namespace
    • resource
  • Check cluster health Number of cluster health check operations.
    Labels:
    • controller_pod
    • controller_namespace
    • name: cluster name.
  • Sync operation Number of sync operations.
    Labels:
    • controller_pod
    • controller_namespace
    • resource
    • result: sync success or failure.
  • Cache sync Number of cache sync operations.
    Labels:
    • controller_pod
    • controller_namespace
    • result: sync success or failure.
  • Controller events Number of events handled by the controller.
    Labels:
    • controller_pod
    • controller_namespace
    • resource
    • operation: includes add, update, delete.