Skip to content

kubernetes

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

kubernetes

README.md

Beats Kubernetes manifests examples

Getting started

This is the list of officially supported Elastic Agents, with example manifests to run them in Kubernetes:

Agent Mode Description
Elastic Agent managed Elastic Agent managed by Fleet setup
Elastic Agent standalone Standalone Elastic Agent setup

Note: kube-state-metrics (KSM) is not installed as part of the above manifests and needs to be installed from kube-state-metrics.

Kustomize Templates

In addition to the above manifests, the list below includes the official kustomize templates to run them in Kubernetes:

Agent Scenario Description
Elastic Agent managed - Default Default Elastic Agent managed by Fleet setup. Kube-state-metrics (KSM) is installed automatically.
Elastic Agent standalone Default Default Standalone Elastic Agent setup. Kube-state-metrics (KSM) is installed automatically.
Elastic Agent managed - With KSM in autosharding configuration Elastic Agent managed by Fleet setup with KMS in autosharding configuration
Elastic Agent standalone - With KSM in autosharding configuration Standalone Elastic Agent setup with KSM in autosharding configuration

How to choose the appropriate scenario:

  • Default scenario, deploys Elastic Agent as Daemonset. The default installation method, that will deploy one agent per Kubernetes node.
  • KSM in autosharding configuration scenario, installs Elastic Agents (that only collect Kube-state-Metrics metrics), along with the KSM as a Side Container. An additional Elastic Agent Leader will be installed as Daemonset and will be responsible for the rest of metrics collection. This scenario is suitable in large Kubernetes deployments. For more information see section below

How to use the kustomize templates

Navigate to the scenario you want to install:

E.g. we want to install the default Elastic Agent Scenario with kustomize.

cd elastic-agent-kustomize/default/elastic-agent-standalone
kubectl apply -k .

Since Kubernetes version 1.14, Kubectl supports the usage of kustomize

Kube State Metrics (KSM) in autosharding configuration

The suggested method of running Elastic Agent in large Kubernetes clusters is to install Elastic Agent as a sidecar container with KSM in order to collect the KSM metrics. As a Kubernetes cluster becomes larger, we need to increase the number of KSM shards and also need to have a dedicated Elastic Agent in order to collect from the specific shard's endpoint. Those are installed as part of a Statefulset.

In more details, the Statefulset acts as a patch of the current KSM Statefuleset pod, meaning that it is responsible to install Elastic Agent with kube-state-metrics datastream configuration as a second container within the same pod. The KSM and Elastic Agent will share the same localhost network to communicate. The Kubernetes provider is disabled (providers.kubernetes.enabled: false).

Apart from the KSM metrics, the rest of the metrics are collected from Elastic Agents deployed as daemonsets

More information about running Elastic Agent along with KSM in sharding mode can be found at elastic-agent-ksm-sharding