Internal alerting

Stakater App Agility Platform also provides fully managed dedicated workload monitoring stack based on Prometheus, Alertmanager and Grafana.

To configure alerting for your application do following:

  1. Create ServiceMonitor for the application
  2. Create AlertmanagerConfig for the application
  3. [Optional] Create PrometheusRule for defining the alerting rule

Note: OpenShift Cluster needs to be on version greater than or equal to 4.7

1. Create ServiceMonitor for the application

Service Monitor uses the service that is used by your application. Then Service Monitor scrapes metrics via that service.

You need to define ServiceMonitor so, the application metrics can be scrapped.

ServiceMonitor can be enabled in Application Chartopen in new window.

.Values.serviceMonitor.enabledEnable ServiceMonitor
.Values.serviceMonitor.endpointsArray of endpoints to be scraped by Prometheus
  enabled: true
  - interval: 5s
    path: /actuator/prometheus
    port: http

2. Create AlertmanagerConfig for the application

You need to define AlertmanagerConfig to direct alerts to your target alerting medium like Slack, PagerDuty, etc.

A sample AlertmanagerConfig can be configured in Application Chartopen in new window.

.Values.alertmanagerConfig.enabledEnable alertmanagerConfig for this app (Will be merged in the base config)
.Values.alertmanagerConfig.spec.routeThe Alertmanager route definition for alerts matching the resource’s namespace. It will be added to the generated Alertmanager configuration as a first-level route
.Values.alertmanagerConfig.spec.receiversList of receivers

We will use Slack as an example here.

Step 1: Create a slack-webhook-config secret which holds Slack webhook URL

kind: Secret
apiVersion: v1
  name: slack-webhook-config
  namespace: <your-namespace>
  webhook-url: <slack-webhook-url-in-base64>
type: Opaque

Step 2: Add a alertmanagerConfig spec to use slack-webhook-config secret created above in step 1, you need to replace <workload-alertmanager-url> with the link of Workload Alertmanager that you can get from Forecastle.

  enabled: true
      receiver: 'slack-webhook'
    - name: 'slack-webhook'
      - apiURL: 
          name: slack-webhook-config
          key: webhook-url
        channel: '#channel-name'
        sendResolved: true
        text: |2-
          {{ range .Alerts }}
          *Alert:* `{{ .Labels.severity | toUpper }}` - {{ .Annotations.summary }}
          *Description:* {{ .Annotations.description }}
            {{ range .Labels.SortedPairs }} *{{ .Name }}:* `{{ .Value }}`
            {{ end }}
          {{ end }}
        title: '[{{ .Status | toUpper }}{{ if eq .Status "firing" }}:{{ .Alerts.Firing | len }}{{ end }}] SAAP Alertmanager Event Notification'
        titleLink: |2
          <workload-alertmanager-url>/#/alerts?receiver={{ .Receiver | urlquery }}
            insecureSkipVerify: true

With this configuration all predefined rules and any new rule which you define should land in the configured slack channel.

Note: AlertmanagerConfig will add a match with your namespace name by default, which will look like this:

        namespace: <your-namespace>

3. [Optional] Create PrometheusRule for the application

Stakater App Agility Platforms comes with lots of Predefined PrometheusRules which covers most of the common use cases.

If required you can definitely create a new PrometheusRule to define for defining alerting rule.

A sample PrometheusRule can be configured in Application Chartopen in new window.

prometheusRule.enabledEnable prometheusRule for this app
prometheusRule.spec.groupsPrometheusRules in their groups to be added
  enabled: true
  groups: []    
  - name: example-app-uptime
    - alert: ExampleAppDown
        message: >-
          The Example App is Down (Test Alert)
      expr: up{namespace="test-app"} == 0
      for: 1m
        severity: critical