KEDA integration

[thisthat]

Goal

Provide integration between KEDA and KeptnMetric

Technical Details

In #601, we introduced a metric adapter that provides metrics via the custom.metrics.k8s.io APIs. KEDA should directly use these APIs to scale a workload.

Acceptance Criteria

• [ ] Adapt the implementation, if necessary

DoD

• [ ] KEDA example is provided

Dependencies

• 603

References

• https://keda.sh/docs/2.7/scalers/metrics-api/

Example

From @thschue:

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: http-scaledobject
  namespace: default
  labels:
    deploymentName: podtato-head-entry
spec:
  maxReplicaCount: 4
  pollingInterval: 3  # Optional. Default: 30 seconds
  cooldownPeriod:  5 # Optional. Default: 300 seconds
  scaleTargetRef:
    name: podtato-head-entry
  triggers:
  - type: metrics-api
    metadata:
      targetValue: "8.8"
      activationTargetValue: "3.8"
      url: "http://keptn-metrics-apiserver.keptn-lifecycle-toolkit-system.svc.cluster.local:9999/api/v1/metrics/default/aggregated-data"
      valueLocation: "value"

[shardulsrivastava]

@thisthat I would like to work on this issue.

[thisthat]

Thank you @shardulsrivastava, I've assigned the ticket to you

[thisthat]

Hey @shardulsrivastava, any progress on this ticket? :)

[Shubh-2402]

Hey @thisthat, I would like to work on this issue. You can assign it to me if it has gone stale. I would need a little guidance to get started though

[thisthat]

Hey @Shubh-2402 thanks for volunteering. The best way to start is by running this example: https://github.com/keptn/lifecycle-toolkit/tree/main/examples/sample-app Here, you define the following Evaluation Definition: https://github.com/keptn/lifecycle-toolkit/blob/main/examples/sample-app/version-3/app-pre-deploy-eval.yaml The idea is to replicate this into a KeptnMetric CRD: https://github.com/keptn/lifecycle-toolkit/blob/main/operator/apis/metrics/v1alpha1/keptnmetric_types.go#L28 The metric operator will then watch this metric from Prometheus continuously. Finally, you should set up KEDA and use the example mentioned in the ticket to create a scaler for the workload. In the URL fields you are using the APIs provided by the operator, their definition is the following: /api/v1/metrics/<namespace>/<metric-name> where <namespace> is the namespace you stored your KeptnMetric CRD and <metric-name> is the name of your KeptnMetric CRD. You can reach out to me directly if you have issue with something :)

[github-actions[bot]]

This issue will be unassigned in 1 week if no further activity is seen. If you are active please provide an update on the status of the issue and if you would like to continue working on it.

[github-actions[bot]]

This issue will be unassigned in 1 week if no further activity is seen. If you are active please provide an update on the status of the issue and if you would like to continue working on it.