Skip to content

Latest commit

聽

History

History
151 lines (104 loc) 路 4.91 KB

README.md

File metadata and controls

151 lines (104 loc) 路 4.91 KB
Raw

Kepler (Kubernetes-based Efficient Power Level Exporter)

Kepler (Kubernetes-based Efficient Power Level Exporter) is a Prometheus exporter. It uses eBPF to probe CPU performance counters and Linux kernel tracepoints.

These data and stats from cgroup and sysfs can then be fed into ML models to estimate energy consumption by Pods.

Setting local kubernetes cluster

  1. Set up Kind cluster.
kind create cluster --config=./01-setup-cluster/01-local-cluster-config.yml
  1. Update the context to use freshly created cluster. (kind does automatically switch context)
kubectl config use-context kind-kepler-demo-local-cluster
kubectl cluster-info --context kind-kepler-demo-local-cluster

Deploying prometheus using helm chart

  1. Setting up Prometheus and Grafana using kube-prometheus-stack.
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update

helm install prometheus prometheus-community/kube-prometheus-stack \
    --namespace monitoring \
    --create-namespace \
    --wait
  1. Verify the resources.
kubectl --namespace monitoring get pods -l "release=prometheus"
#NAME                                                   READY   STATUS    RESTARTS   AGE
#prometheus-kube-prometheus-operator-8484fbd69d-g7lvl   1/1     Running   0          4m3s
#prometheus-kube-state-metrics-6f659dc9cb-xjzx2         1/1     Running   0          4m3s
#prometheus-prometheus-node-exporter-g7mg2              1/1     Running   0          4m3s
  1. And, we have alert manager as well as grafana also running.
kubectl --namespace monitoring get pods
NAME                                                     READY   STATUS    RESTARTS   AGE
alertmanager-prometheus-kube-prometheus-alertmanager-0   2/2     Running   0          14h
prometheus-grafana-6ffd599f89-wwrf5                      3/3     Running   0          14h
prometheus-kube-prometheus-operator-8484fbd69d-g7lvl     1/1     Running   0          14h
prometheus-kube-state-metrics-6f659dc9cb-xjzx2           1/1     Running   0          14h
prometheus-prometheus-kube-prometheus-prometheus-0       2/2     Running   0          14h
prometheus-prometheus-node-exporter-g7mg2                1/1     Running   0          14h
  1. Doing port forward for grafana to access the dashboard.

It's easier to configure using k9s. Start the utility and type namespace to list all namespaces and choose monitoring.

list-namespaces

Choose the pod with prometheus-grafana-* and press shift-f for port forwarding.

port-forward-3002

I have used port 3002 as the port 3000 in my machine is already occupied by different service.

  1. Access the grafana dashboard on url: Grafana local and use the credentials: username as admin and password as prom-operator.

Deploying kepler using helm chart

  1. Setting up kepler repo and installing latest version.
helm repo add kepler https://sustainable-computing-io.github.io/kepler-helm-chart
helm repo update
  1. Find the latest version of the kepler.
helm search repo kepler
  1. Make a dry run.
helm install kepler kepler/kepler --namespace kepler --create-namespace --dry-run --devel
  1. Deploy kepler
helm install kepler kepler/kepler \
    --namespace kepler \
    --create-namespace \
    --set serviceMonitor.enabled=true \
    --set serviceMonitor.labels.release=prometheus
  1. Verify the installation
KPLR_POD=$(
    kubectl get pod \
        -l app.kubernetes.io/name=kepler \
        -o jsonpath="{.items[0].metadata.name}" \
        -n kepler
)
kubectl wait --for=condition=Ready pod $KPLR_POD --timeout=-1s -n kepler

Accessing the Kepler Exporter dashboard

As there is no load and consumption, hence the load will almost equivalent to zero. Go to Dashboard and click on new and choose Import. With import option, we can add predefined JSON files from the kepler github repo. The json files also present in folder 03-grafana-dashboard.

And, then simply upload the JSON files, and access it.

grafana-kepler-dashboard

How does it help?

With help of the dashboard you can monitor the resources, and can optimize the usage.

How to unistall?

  1. List all the deployments.
$ helm list --all-namespaces | awk '{print $1, $2}'
NAME            NAMESPACE
kepler          kepler
prometheus      monitoring
  1. Uninstall the deployment.
$ helm uninstall kepler --namespace kepler
release "kepler" uninstalled

$ helm uninstall prometheus --namespace monitoring
release "prometheus" uninstalled

Thank you

Thank you for contributing towards the green development 馃實.