This repository has been archived by the owner. It is now read-only.
87 lines (68 sloc) 4.48 KB

Configuring sources

Heapster can get data from multiple sources (although at this moment we support only one kind - Kubernetes). They are specified in the command line via the --source flag. The flag takes an argument of the form PREFIX:CONFIG[?OPTIONS]. Options (optional!) are specified as URL query parameters, separated by & as normal. This allows each source to have custom configuration passed to it without needing to continually add new flags to Heapster as new sources are added. This also means Heapster can capture metrics from multiple sources at once, potentially even multiple Kubernetes clusters.

Current sources


To use the kubernetes source add the following flag:


If you're running Heapster in a Kubernetes pod you can use the following flag:


Heapster requires an authentication token to connect with the apiserver securely. By default, Heapster will use the inClusterConfig system to configure the secure connection. This requires Kubernetes version v1.0.3 or higher and a couple extra Kubernetes configuration steps. Firstly, for your apiserver you must create an SSL certificate pair with a SAN that includes the ClusterIP of the Kubernetes service. Look here for an example of how to properly generate certs. Secondly, you need to pass the ca.crt that you generated to the --root-ca-file option of the controller-manager. This will distribute the root CA to /var/run/secrets/ of all pods. If you are using ABAC authorization (as opposed to AllowAll which is the default), you will also need to give the system:serviceaccount:<namespace-of-heapster>:default readonly access to the cluster (look here for more info).

If you don't want to setup inClusterConfig, you can still use Heapster! To run without auth, use the following config:


This requires the apiserver to be setup completely without auth, which can be done by binding the insecure port to all interfaces (see the apiserver --insecure-bind-address option) but WARNING be aware of the security repercussions. Only do this if you trust EVERYONE on your network.

Note: Remove "monitoring-token" volume from heapster controller config if you are running without auth.

Alternatively, you can use a heapster-only serviceaccount like this:

cat <<EOF | kubectl create -f -
apiVersion: v1
kind: ServiceAccount
  name: heapster

This will generate a token on the API server. You will then need to reference the service account in your Heapster pod spec like this:

apiVersion: "v1"
kind: "ReplicationController"
    name: "heapster"
  name: "monitoring-heapster-controller"
  replicas: 1
    name: "heapster"
        name: "heapster"
      serviceAccount: "heapster"
          image: "kubernetes/heapster:v0.13.0"
          name: "heapster"
            - "/heapster"
            - "--source=kubernetes:http://kubernetes-ro?inClusterConfig=false&useServiceAccount=true&auth="
            - "--sink=influxdb:http://monitoring-influxdb:80"

This will mount the generated token at /var/run/secrets/ in the Heapster container.

The following options are available:

  • inClusterConfig - Use kube config in service accounts associated with Heapster's namespace. (default: true)
  • kubeletPort - kubelet port to use (default: 10255)
  • kubeletHttps - whether to use https to connect to kubelets (default: false)
  • insecure - whether to trust Kubernetes certificates (default: false)
  • auth - client auth file to use. Set auth if the service accounts are not usable.
  • useServiceAccount - whether to use the service account token if one is mounted at /var/run/secrets/ (default: false)

There is also a sub-source for metrics - kubernetes.summary_api - that uses a slightly different, memory-efficient API for passing data from Kubelet/cAdvisor to Heapster. It supports the same set of options as kubernetes. Sample usage:

 - --source=kubernetes.summary_api:''