AI Driven DevOps++

Greetings to All!!

Technologies are changing very rapidly and AI / ML is getting into every space, in fact even into our Houses...) Why cant we let it in to the Engineering activities, Software Development Life-cycle, Business process Management, Chaos Engineering, Quality Assurance, Operations Management and Support, Service Reliability Engineering etc. Well, I would like to confine this post to AI Driven DevOps++. I hope you remember, we have built a platform called DevOps++ , which is a next generation platform with lot of features (please refer my previous post for more details). This platform will have more extended features in near future, of which AI Driven DevOps would be the critical and my favorite one. The critical uses cases that can be part of this are:

1. AI Driven Environment Setup: Let's say you have bagged a new project and you enter all the details of the project, customer information, estimation details (resource loading, efforts etc) into the enterprise CRM system or any other system, before you hand over it to a delivery team. This would be entry point for the AI Agent. With minimal set of configuration items as inputs to the AI Agent, it would automatically set up a development environment for your engineering teams either on Cloud or on-premise data center. It traverses through the projects data store across the company and identify the projects that are similar to the new one, and extract the information, analyzes the data and provide you the output in the form of documents related Infrastructure, software to be used, Configuration Management system model, Indicative Test plans & Test Cases, and a project plan. The environment set by this AI Agent would include, a Configuration management system (GIT or any other..), Development environment with the required IDE (VS or Eclipse etc), Database (NoSQL or MySQL etc... ) and automated CI-CD framework, and Code Quality plugins for Static code analysis, finding vulnerabilities in code etc. In short, you are ready with environment setup even before you on-board your teams. This would save lot of time. As you understand AI and ML depends on the DATA, and the output depends on the Quality and the size of the Data that you have in the organization.
2. Code Quality: ML models are created to continuously learn the different patterns related to Static Code Analysis and Vulnerabilities and applied automatically to the new code. This would be checked while checking - in code to GIT or as part of build and deployment process. Your peer reviews / code reviews process would be completely automated.

The other critical use cases are:

1. Automated Test Cases, test case optimization
2. Operational Support - Self Healing options for both applications and Infrastructure (This areas is already maturing ...)
3. Service Reliability through Chaos Engineering and many more.

As i said earlier, the quality of output depends on the quality of the data that you have in the systems that will be used for creating patterns and ML models.

Sounds interesting ...right?

Watch out this space.. More to come soon ...

-- Regards

Murali Mohan Josyula (JMM)