custom-metrics-apiserver

Custom Metrics Adapter Server Boilerplate

Purpose

This repository contains boilerplate code for setting up an implementation of the Metrics APIs:

• Custom metrics (k8s.io/metrics/pkg/apis/custom_metrics)
• External metrics (k8s.io/metrics/pkg/apis/external_metrics)

It includes the necessary boilerplate for setting up an implementation (generic API server setup, registration of resources, etc), plus an implementation for testing that allows setting custom metric values over HTTP.

How to use this repository

This repository is designed to be used as a library. First, implement one or more of the metrics provider interfaces in pkg/provider (for example, CustomMetricsProvider), depending on which APIs you want to support.

Then, use the AdapterBase in pkg/cmd to initialize the necessary flags and set up the API server, passing in your providers.

More information can be found in the getting started guide, and the testing implementation can be found in the test-adapter directory.

Prerequisites

Go: this library requires the same version of Go as Kubernetes.

Test Adapter

There is a test adapter in this repository that can be used for testing changes to the repository, as a mock implementation of the APIs for automated unit tests, and also as an example implementation.

Note that this adapter should not be used for production. It's for writing automated e2e tests and serving as a sample only.

To build and deploy it:

# build the test-adapter container as $REGISTRY/k8s-test-metrics-adapter-amd64
export REGISTRY=<some-prefix>
make test-adapter-container

# push the container up to a registry (optional if your cluster is local)
docker push $REGISTRY/k8s-test-metrics-adapter-amd64

# launch the adapter using the test adapter deployment manifest
kubectl apply -f test-adapter-deploy/testing-adapter.yaml

When the deployment is ready, you can define new metrics on the test adapter by querying the write endpoint:

# set up a proxy to the API server so we can access write endpoints
# of the testing adapter directly
kubectl proxy &
# write a sample metric -- the write paths match the same URL structure
# as the read paths, but at the /write-metrics base path.
# data needs to be in json, so we also need to set the content-type header
curl -X POST \
  -H 'Content-Type: application/json' \
  http://localhost:8001/api/v1/namespaces/custom-metrics/services/custom-metrics-apiserver:http/proxy/write-metrics/namespaces/default/services/kubernetes/test-metric \
  --data-raw '"300m"'

# you can pipe to `jq .` to pretty-print the output, if it's installed
# (otherwise, it's not necessary)
kubectl get --raw "/apis/custom.metrics.k8s.io/v1beta2" | jq .
# fetching certain custom metrics of namespaced resources
kubectl get --raw "/apis/custom.metrics.k8s.io/v1beta2/namespaces/default/services/kubernetes/test-metric" | jq .

If you wanted to target a simple nginx-deployment and then use this as an HPA scaler metric, something like this would work following the previous curl command:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: nginx-deployment
  namespace: default
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: nginx-deployment
  minReplicas: 1
  maxReplicas: 10
  metrics:
  - type: Object
    object:
      metric:
        name: test-metric
      describedObject:
        apiVersion: v1
        kind: Service
        name: kubernetes
      target:
        type: Value
        value: 300m

You can also query the external metrics:

kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1" | jq .
# fetching certain custom metrics of namespaced resources
kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1/namespaces/default/my-external-metric" | jq .

Compatibility

The APIs in this repository follow the standard guarantees for Kubernetes APIs, and will follow Kubernetes releases.

Community, discussion, contribution, and support

Learn how to engage with the Kubernetes community on the community page.

You can reach the maintainers of this repository at:

• Slack: channel #sig-instrumentation
• Mailing List

Code of Conduct

Participation in the Kubernetes community is governed by the Kubernetes Code of Conduct.

Contribution Guidelines

See CONTRIBUTING.md for more information.