Skip to main content

Kubernetes

Overview

Container orchestrators such as Kubernetes and Amazon ECS provide native capabilities for deploying and managing the lifecycles of Sensors, such as deploying, recycling, high-available and version management.

Components

  • Operator
  • Sensor
  • Cluster Manager

To learn more about the Upwind components and their architecture review Upwind Components

Prerequisites

Minimum Kubernetes Versions

The Kubernetes project maintains release branches for the most recent three minor releases, review current supported releases .

Upwind will maintain support for all maintained releases , older releases are not tested and not recommended.

The following Kubernetes versions:

  1. Kubernetes - Versions 1.20 and above.
  2. OpenShift - Versions 4.6 and above.
  3. Amazon EKS - Versions 1.20 and above.
  4. Google Kubernetes Engine (GKE) - Versions 1.20 and above.
  5. Azure Kubernetes Engine (AKS) - Versions 1.20 and above.

Container Runtimes

  1. Docker Engine
  2. Containerd
  3. CRI-O

Minimum Docker Engine Versions

Upwind will maintain support for all maintained releases

Minimum Containerd Versions

Upwind will maintain support for all maintained releases

Minimum CRI-O Versions

CRI-O also follows the Kubernetes n-2 release version. If a Kubernetes release goes End of Life, then the corresponding CRI-O version can be considered in the same way.

Supported Kubernetes types

  • ReplicaSet
  • Deployment
  • StatefulSet
  • Standalone pods
  • Job
  • CronJob
  • DaemonSet
  • Kubernetes Nodes

Installation

The installation choices available for deploying the Upwind Sensor on Kubernetes are as follows:

Upwind Operator

Here are a few benefits of installing Upwind Sensor using an Operator:

  • The Operator has built-in default configurations for memory & CPU consumption based on Upwind best practices.
  • The Operator configuration can automatically upgrade the running Upwind Sensors once a new version becomes available.
  • It is advisable to avoid manual DaemonSet configuration due to the substantial potential for errors.