Monitor your app

Metrics

You can export your application’s metrics by creating a ServiceMonitor that watches your Service. The ServiceMonitor observes all services that match the specified selector. For example, the following ServiceMonitor monitors all services with the label app: sampleapp.

servicemonitor.yaml

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: sample-servicemonitor
spec:
  selector:
    matchLabels:
      app: sampleapp
  endpoints:
  - port: http

Tracing

Tracing is enabled with Tempo.

This tool will be able to receive all the sent traces and supports OTEL (OpenTelemetry). To send your traces, direct them to tempo.k8saas-system.svc.cluster.local.:4317. Traces can be found in Grafana. Use “Explore”, then find “Traces”.

Logs

Application logs are collected by default using the Logging Operator. You do not need additional configuration to collect your application’s logs, as all application logs are automatically captured. Logs can be found in Grafana. On “Dashoards”, find the loki/logs one.

Custom dashboard

To provision a Grafana Dashboard using a declarative way, create a ConfigMap with the grafana_dashboard: "1" label containing the dashboard JSON manifest.

dashboard.yaml

apiVersion: v1
kind: ConfigMap
metadata:
  name: sampleapp-dashboard
  labels:
    grafana_dashboard: "1"
data:
  sampleapp-dashboard.json: |
    { ... }

Predefined Metrics

Some metrics are exposed by default, such as certificate status and ingress requests. For a detailed example of a dashboard exposing various metrics, check out this dashboard YAML example.

Alerting

By creating a PrometheusRule, you can create alerting rules when your applications meets your desired requirements. This alert is fired if the there is less than 2 ready replicas for the “application-api” deployment.

Make sure to create an according AlertmanagerConfig in order to let fire alerts based on this rule.

rule.yaml

apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
  name: application-api-down-rule
  labels:
    release: monitoring
spec:
  groups:
    - name: "application.rules"
      rules:
        - alert: ApplicationAPIDown
          for: 1m
          expr: kube_deployment_status_replicas_ready{deployment="application-api"} < 2
          labels:
            severity: critical
            app: application-api
          annotations:
            message: The API deployment has less than 2 pods running.
            summary: the current deployment "application-api" need to have at least 2 replicas running

Validate integration

To validate your monitoring integration, navigate to the Grafana UI and locate the relevant dashboard. Ensure the selected data aligns with your monitoring configuration and is being successfully scraped.