k8sPulse

k8sPulse Documentation

Overview

k8sPulse is an open-source Kubernetes monitoring tool designed for emergency scenarios when you only have kubectl access. It leverages the Python Kubernetes native client and OpenAI to provide intelligent recommendations on cluster health.

Installation Guide

Prerequisites

To use k8sPulse, you’ll need pipx installed to easily manage the tool as an isolated package.

Installing k8sPulse with pipx

To install k8sPulse using pipx, run the following command:

pipx install git+https://github.com/raestrada/k8sPulse.git@v0.7.0

This command installs the latest tagged version (v0.7.0) of k8sPulse.

Installing pipx

On Linux and macOS

To install pipx on Linux or macOS, run:

python3 -m pip install --user pipx
python3 -m pipx ensurepath

Ensure python3 and pip are installed on your system.

On Windows

To install pipx on Windows, you can use the following command in PowerShell:

python -m pip install --user pipx
python -m pipx ensurepath

After installing pipx, close and reopen your terminal or run refreshenv if you’re using the Windows Command Prompt.

Usage

Once k8sPulse is installed, you can use the following commands and options to generate reports and monitor your Kubernetes cluster:

Generating a Report

To generate a report with default settings, use:

k8spulse

This command will generate a Kubernetes cluster report for the default staging environment.

Available Options

• --env-name
  • Description: Specify the environment name for the report.
  • Default Value: staging

  • Usage:

    k8spulse --env-name production
    

• --interval
  • Description: Set the interval (in seconds) between report generations.
  • Default Value: 300 (5 minutes)

  • Usage:

    k8spulse --interval 600
    

• --use-ai
  • Description: Use OpenAI to generate recommendations based on the report.

  • Usage:

    k8spulse --use-ai
    

• --git-commit
  • Description: Automatically commit and push the generated report to the Git repository.

  • Usage:

    k8spulse --git-commit
    

• --gpt-model
  • Description: Specify which GPT model to use for recommendations.
  • Default Value: gpt-4o

  • Usage:

    k8spulse --gpt-model gpt-4o
    

• --zombies
  Description: Enable experimental zombie process detection in non-running pods using a Bash script and kubectl. This approach provides better performance compared to using the Python Kubernetes API client.
  Usage:

  k8spulse --zombies
  

  Configuring Cast.AI Events Tracking

To enable tracking of Cast.AI events in your k8sPulse reports:

1. Ensure you have a CAST AI API key set as an environment variable:

   export CAST_AI_API_KEY=your_cast_ai_api_key_here
   

2. Set the cluster ID you want to track by exporting it as an environment variable:

   export CAST_AI_CLUSTER_ID=your_cluster_id_here
   

3. The Cast.AI events will be automatically included in your k8sPulse report, providing details about node additions, deletions, and autoscaler actions.

Example

Below is an example of using multiple options together:

k8spulse --env-name production --interval 600 --use-ai --git-commit --gpt-model got-4o

This command generates a report for the production environment every 10 minutes, uses OpenAI for recommendations, commits the report to Git, and uses the got-4o GPT model.

Generating the HTML Report

After generating a report, you can view it by opening the generated staging_statistics.html file in your browser. The report provides a visual overview of the Kubernetes cluster, including gauges, unusual events, and detailed insights.

Contributing

We welcome contributions to k8sPulse! To get started, visit our GitHub repository and check out the contributing guidelines.

License

k8sPulse is open source and available under the MIT License.

This site is open source. Improve this page.