Run Metricbeat on Kubernetes
editRun Metricbeat on Kubernetes
editYou can use Metricbeat Docker images on Kubernetes to retrieve cluster metrics.
Running Elastic Cloud on Kubernetes? See Run Beats on ECK.
Kubernetes deploy manifests
editYou deploy Metricbeat as a DaemonSet to ensure that there’s a running instance on each node of the cluster. These instances are used to retrieve most metrics from the host, such as system metrics, Docker stats, and metrics from all the services running on top of Kubernetes.
In addition, one of the Pods in the DaemonSet will constantly hold a leader lock which makes it responsible for handling cluster-wide monitoring. This instance is used to retrieve metrics that are unique for the whole cluster, such as Kubernetes events or kube-state-metrics. You can find more information about leader election configuration options at Autodiscover.
Note: If you are upgrading from older versions, please make sure there are no redundant parts as left-overs from the old manifests. Deployment specification and its ConfigMaps might be the case.
Everything is deployed under the kube-system
namespace by default. To change
the namespace, modify the manifest file.
To download the manifest file, run:
curl -L -O https://raw.githubusercontent.com/elastic/beats/8.4/deploy/kubernetes/metricbeat-kubernetes.yaml
If you are using Kubernetes 1.7 or earlier: Metricbeat uses a hostPath volume to persist internal data. It’s located
under /var/lib/metricbeat-data
. The manifest uses folder autocreation (DirectoryOrCreate
), which was introduced in
Kubernetes 1.8. You need to remove type: DirectoryOrCreate
from the manifest and create the host folder yourself.
Settings
editBy default, Metricbeat sends events to an existing Elasticsearch deployment, if present. To specify a different destination, change the following parameters in the manifest file:
- name: ELASTICSEARCH_HOST value: elasticsearch - name: ELASTICSEARCH_PORT value: "9200" - name: ELASTICSEARCH_USERNAME value: elastic - name: ELASTICSEARCH_PASSWORD value: changeme
Running Metricbeat on master nodes
editKubernetes master nodes can use taints to limit the workloads that can run on them. To run Metricbeat on master nodes you may need to update the Daemonset spec to include proper tolerations:
spec: tolerations: - key: node-role.kubernetes.io/master effect: NoSchedule
Red Hat OpenShift configuration
editIf you are using Red Hat OpenShift, you need to specify additional settings in
the manifest file and grant the metricbeat
service account access to the privileged SCC:
-
In the manifest file, edit the
metricbeat-daemonset-modules
ConfigMap, and specify the following settings underkubernetes.yml
in thedata
section:kubernetes.yml: |- - module: kubernetes metricsets: - node - system - pod - container - volume period: 10s host: ${NODE_NAME} hosts: ["https://${NODE_NAME}:10250"] bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token ssl.certificate_authorities: - /path/to/kubelet-service-ca.crt - module: kubernetes metricsets: - proxy period: 10s host: ${NODE_NAME} hosts: ["localhost:29101"]
kubelet-service-ca.crt
can be any CA bundle that contains the issuer of the certificate used in the Kubelet API. According to each specific installation of Openshift this can be found either insecrets
or inconfigmaps
. In some installations it can be available as part of the service account secret, in/var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt
. In case of using Openshift installer for GCP then the followingconfigmap
can be mounted in Metricbeat Pod and useca-bundle.crt
inssl.certificate_authorities
:Name: kubelet-serving-ca Namespace: openshift-kube-apiserver Labels: <none> Annotations: <none> Data ==== ca-bundle.crt:
-
If
https
is used to accesskube-state-metrics
, add the following settings to themetricbeat-daemonset-config
ConfigMap under the kubernetes autodiscover configuration for thestate_*
metricsets:bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token ssl.certificate_authorities: - /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt
-
Grant the
metricbeat
service account access to the privileged SCC:oc adm policy add-scc-to-user privileged system:serviceaccount:kube-system:metricbeat
This command enables the container to be privileged as an administrator for OpenShift.
-
If the namespace where elastic-agent is running has the
"openshift.io/node-selector"
annotation set, elastic-agent might not run on all nodes. In this case consider overriding the node selector for the namespace to allow scheduling on any node:oc patch namespace kube-system -p \ '{"metadata": {"annotations": {"openshift.io/node-selector": ""}}}'
This command sets the node selector for the project to an empty string. If you don’t run this command, the default node selector will skip master nodes.
for openshift versions prior to the version 4.x additionally you need to modify the DaemonSet
container spec in the manifest file to enable the container to run as privileged:
securityContext: runAsUser: 0 privileged: true
Load Kibana dashboards
editMetricbeat comes packaged with various pre-built Kibana dashboards that you can use to visualize metrics about your Kubernetes environment.
If these dashboards are not already loaded into Kibana, you must install Metricbeat
on any system that can connect to the Elastic Stack, and then run the setup
command to load the dashboards. To learn how,
see Load Kibana dashboards.
If you are using a different output other than Elasticsearch, such as Logstash, you need to Load the index template manually and Load Kibana dashboards.
Deploy
editMetricbeat gets some metrics from kube-state-metrics.
If kube-state-metrics
is not already running, deploy it now (see the
Kubernetes
deployment docs).
To deploy Metricbeat to Kubernetes, run:
kubectl create -f metricbeat-kubernetes.yaml
To check the status, run:
$ kubectl --namespace=kube-system get ds/metricbeat NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE-SELECTOR AGE metricbeat 32 32 0 32 0 <none> 1m
Metrics should start flowing to Elasticsearch.
Deploying Metricbeat to collect cluster-level metrics in large clusters
editThe size and the number of nodes in a Kubernetes cluster can be fairly large at times, and in such cases the Pod that will be collecting cluster level metrics might face performance issues due to resources limitations. In this case users might consider to avoid using the leader election strategy and instead run a dedicated, standalone Metricbeat instance using a Deployment in addition to the DaemonSet.