Harness today previewed two additional modules for its artificial intelligence (AI) platform for DevOps that automate code maintenance and automatically rollback a deployment whenever an issue is detected.
Additionally, Harness is adding an Architect Mode to the Harness AI platform that interactively engages DevOps engineers to design pipelines based on best practices for security, quality and compliance that have been defined by the organization, in addition to making available an AI agent for its internal developer portal (IDP).
Finally, Harness Release Orchestration has been extended using an AI capability that makes it possible to now model new release processes using natural language, which Harness then translates into structured YAML code.
Company Field CTO Nick Durkin said these latest editions to the Harness AI platform is part of an ongoing effort to leverage AI to both reduce the overall amount of toil and the need to depend on DevOps specialists to perform rote tasks.
For example, the Harness Autonomous Code Maintenance (ACM) module makes it possible to describe a task in plain English to create a software build that has passed security and functional tests. If a build failure happens after a developer opens a pull request, ACM will also repeatedly analyze the issue, generate a potential fix, create a new branch, and submit a pull request with the suggested remediation to self-heal the build until it is successfully completed. Additionally, it will detect stale feature flags, remove the references, and open pull requests for developers to approve.
The AI Verification and Rollback module, meanwhile, provides access to an AI agent that connects the observability platform a DevOps team has adopted. Armed with those insights, it can then discover relevant metrics and log queries for the service you’re deploying to build a comprehensive health verification profile. If it detects a problem, the AI agent will trigger an automatic rollback to the last known good version of the build.
It’s not clear how DevOps teams will be structured in the age of AI, but the one clear thing is that there will be less dependency on specialists, said Durkin. Instead of having to answer a continuous stream of questions from colleagues, DevOps specialists will be able to spend more time resolving issues that will enable organizations to deploy more complex applications at scale, he added.
At the same time, more junior developers will have an opportunity to grow faster as the level of expertise required to successfully build and deploy software is sharply reduced, he added.
Overall, DevOps teams are still spending 60% to 70% of their time on trivial tasks that few of them enjoy, said Durkin. AI will enable DevOps engineers to focus much more of their time on tasks they are actually passionate about versus all the minutiae they currently have to manage, he added.
Naturally, it’s still early days so far as the adoption of AI in DevOps workflows is concerned, but the one thing that is already apparent is that many of the tedious tasks that today conspire to burn out DevOps teams are soon going to be eliminated.
