https://github.com/int128/spot-handler

Kubernetes controller for observability of Amazon EC2 Spot Interruption
https://github.com/int128/spot-handler

aws kubebuilder kubernetes kubernetes-controller observability spot-instances

Last synced: 7 months ago
JSON representation

Kubernetes controller for observability of Amazon EC2 Spot Interruption

• Host: GitHub
• URL: https://github.com/int128/spot-handler
• Owner: int128
• Created: 2024-07-07T05:59:54.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-03-14T15:22:04.000Z (7 months ago)
• Last Synced: 2025-03-14T16:29:45.075Z (7 months ago)
• Topics: aws, kubebuilder, kubernetes, kubernetes-controller, observability, spot-instances
• Language: Go
• Homepage:
• Size: 489 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# spot-handler [![go](https://github.com/int128/spot-handler/actions/workflows/go.yaml/badge.svg)](https://github.com/int128/spot-handler/actions/workflows/go.yaml)

This is a Kubernetes controller for the observability of spot interruptions.

It provides the Kubernetes events for the [Amazon EC2 Spot Instance interruption notices](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/spot-instance-termination-notices.html).

## Features

### Observability of spot interruptions

When a spot interruption occurs, it may cause a problem such as network error.

It is very important to measure the number of affected nodes or pods.

This controller provides the following events in Kubernetes:

- `SpotInterrupted` event on `Node` resource

- `SpotInterrupted` event on `Pod` resource

### Pod termination handler

When a spot interruption occurs, possibly a pod is not gracefully terminated.

For example, a pod disruption budget (PDB) may prevent the pod from being deleted.

This controller can delete the affected pods to initiate the graceful termination.

## Getting Started

You need to set up an EventBridge rule and an SQS queue to receive the spot interruption messages.

```mermaid

graph LR

  EventBridge --> SQS --> Controller --> Event --> Datadog

  subgraph AWS

    EventBridge

    SQS

  end

  subgraph Kubernetes cluster

    Controller

    Event

  end

```

Deploy the manifest from the GitHub Releases.

Configure the Queue resource with the SQS URL.

```yaml

apiVersion: spothandler.int128.github.io/v1

kind: Queue

metadata:

  name: eks-spot-handler

spec:

  url: https://sqs.us-east-2.amazonaws.com/123456789012/eks-spot-handler

```