You created a pipeline that can deploy your source code changes to your infrastructure in instance groups for self-healing. One of the changes negatively affects your key performance indicator. You are not sure how to fix it, and investigation could take up to a week. What should you do?

Reverting the source code change and rerunning the deployment pipeline is the most efficient and reliable way to mitigate the negative impact. This leverages your existing CI/CD pipeline, ensuring a consistent and automated rollback to a known good state. Logging in to a server and iterating locally is not scalable, introduces manual errors, and bypasses your established deployment process. Logging into servers to manually swap code is also error-prone, not scalable, and circumvents the pipeline. Changing the instance group template and deleting all instances would work, but it's a more disruptive and potentially slower process than simply reverting the source code and letting the pipeline redeploy. The pipeline is designed to handle such changes efficiently.