How to rollback a Python application with Kubernetes?

Table of Contents

• Introduction
• Checking Deployment History in Kubernetes
  • Step 1: View Deployment History
  • Step 2: View Detailed History of a Specific Revision
• Rolling Back a Python Application
  • Step 3: Roll Back the Deployment
  • Step 4: Verify the Rollback
• Managing Rollbacks in Kubernetes
  • Step 5: Set Maximum Rollout Revisions
  • Step 6: Pause and Resume Rollouts
• Conclusion

Introduction

Rolling back a Python application in Kubernetes involves reverting the application to a previous deployment version. Kubernetes stores the history of deployments, allowing you to easily roll back in case of issues with a newer version. This guide explains how to use Kubernetes commands to perform a rollback and manage your Python application's deployment history.

Checking Deployment History in Kubernetes

Before initiating a rollback, it's important to check the deployment history to confirm which version you want to roll back to.

Step 1: View Deployment History

You can list the history of your deployments to see the different revisions:

This command will show the revision numbers and details of each deployment.

Example output:

Each revision corresponds to a specific deployment change, with the most recent one being the current version. The CHANGE-CAUSE provides insight into what triggered the deployment (if set).

Step 2: View Detailed History of a Specific Revision

To get more details about a specific revision, use the following command:

This command provides information about the containers and the image version used in that specific revision, helping you verify the deployment before rolling back.

Rolling Back a Python Application

Once you've confirmed the revision you want to roll back to, you can perform the rollback with a simple command.

Step 3: Roll Back the Deployment

To roll back your Python application to the previous version, use the following command:

This command rolls back to the last successful deployment. If you need to roll back to a specific revision, you can specify it:

This command will roll back the deployment to the version associated with revision 1.

Step 4: Verify the Rollback

After performing the rollback, you should verify that the deployment has successfully reverted to the desired state:

This command will show the status of the rollback and ensure that the pods are being deployed with the correct version.

Managing Rollbacks in Kubernetes

Kubernetes provides additional controls to manage the rollback process and ensure application stability.

Step 5: Set Maximum Rollout Revisions

By default, Kubernetes stores the last 10 deployment revisions. You can configure the number of revisions to retain using the revisionHistoryLimit field in the deployment YAML:

This will limit the number of revisions Kubernetes stores, helping to manage memory and storage.

Step 6: Pause and Resume Rollouts

If you need to make changes to the deployment and ensure everything is correct before continuing, you can pause the rollout:

Once the changes are made and verified, you can resume the rollout:

Pausing and resuming gives you more control over the deployment process, especially during upgrades or troubleshooting.

Conclusion

Rolling back a Python application in Kubernetes is straightforward using the kubectl rollout undo command. With Kubernetes' built-in deployment history, you can easily manage and revert deployments, ensuring application stability. This process is crucial for maintaining the smooth operation of your Python applications in production environments, especially when encountering issues with new updates.