Overview

“Automation is crucial to warrant a smooth change and release management process”

Changes to platform foundational services such as IAM, network or firewalling, can have a substantial impact to solutions and services that rely on those services. ​

High impact changes must be assessed and tested carefully before implementation . This is of course also true for any solution. Automation makes sure these changes can be executed quickly and reliably without sacrificing security or quality. ​

Automation is critical to avoid human error when deploying changes to production, after they have been tested and accepted. At all times, ensure that changes can be reversed by keeping a template of the last working version. ​

Using automation pipelines, resources in test/acceptance environments can be automatically created, tested and destroyed, which makes testing and acceptance extremely time and cost effective. Similarly, automation pipelines can be used particularly well to aim for continuous compliance through safeguards in the pipeline.​

Updates and changes to services can originate from the cloud provider. Therefore, messages from the provider should be checked daily and, if needed, action taken to avoid interruptions in the services. It is a good idea to give DevOps teams a heads up about upcoming changes from the cloud provider that may have impact.​

Due to the rapid pace of changes the original change process will no longer be suitable. The change advisory board will not be able to gather for each and every change. Combine this with more autonomous teams and it is clear the change process will need to be revised.​

The use of approved/standardised service blueprints or configurations helps to speed up the change and release cadence for the teams. Full automation of the change and release process will increase efficiency and reduce potential risks.

Activities checklist

Initial:

• Creating a list of change types and potential impact​
• Making automation of changes the standard way of working​
• Considering making automation the only way to change production environments

Recurring:

• Testing changes with an automation pipeline​
• Monitoring the messages from the cloud partner regarding upcoming changes impacting the platform​
• Informing DevOps teams about (platform) changes

RASCI