DevOps company Harness has introduced new services including AI Assistants for DevOps, coding and QA; cloud development environments (CDEs) for cloud-based coding; an Artifact Registry; Database DevOps; and supply-chain security. AI features are primarily delivered by Google Cloud using its Gemini service.
The new features were shown at the company’s “unscripted” event in London and online.
CDEs, also known as Gitspaces, are now in public beta, with the promise of preconfigured developer tools and libraries, consistent environments across a team, and keeping source code away from local PCs. CDEs are configured using the devcontainer.json specification created by Microsoft.
Harness states that developers can use their IDE of choice, but currently that means either VS Code desktop or VS Code in the browser. IntelliJ IDEA and others are to follow, we were told. The enviroments can be spun up in either US East or US West; the cloud provider is not stated but might be Google Cloud Platform (GCP), given that Harness CEO Jyoti Bansai spoke at the event of a partnership with GCP, including use of Gemini and of Google Cloud Build, and that Google Cloud was a co-sponsor of the conference.
There is version of Harness that is open source on GitHub under the Apache 2.0 license and can be used for self-hosting. A self-hosted option for Gitspaces is planned.
At the event, Bansai also enthused about the benefits of AI for DevOps, stating that “AI has to play a major role in how software engineering is done”. He spoke about an “agentic workflow,” meaning that AI-driven agents will suggest and automate DevOps tasks, and respond to plain English commands.
AI Assistants will include DevOps, coding, and QA, now in preview, and coming in future, FinOps and AppSec (application security).
The Harness platform is centred on continuous integration pipelines with a visual user interface and an underlying YAML definition. In an example at the conference, the DevOps engineer types “add sast scans”, where sast is short for static application security testing, and the DevOps Assistant proposed changes to the YAML adding Snyk, Sonarqube and Wiz stages to the pipeline.
In QA, Harness described what Bansai called “intent-based testing” where the developer describes the intention of the test and the AI generates the code, using UI testing as an example.
The AI Code Assistant is primarily based on Google Gemini though OpenAI was also mentioned. Features include code completion, function generation, code refactoring and debugging, code explanation, search using natural language, and pull request generation.
Another new module now in beta is Database DevOps, which includes database changes in the CI pipeline. The idea is that developers can “treat database code like application code” with versioning, git repository check-in, rollback when needed, and automated deployment. Initially supported database managers are Oracle, Microsoft SQL Server, MySQL, MongoDB, and PostgreSQL.
Harness also previewed the Harness Artifact Registry, which includes security features to identify vulnerabilities.
This is a significant array of new features for the Harness platform, though one thing not discussed at the event was pricing. Further, most of the new modules (with Database DevOps perhaps an exception) repackage features already available elsewhere. The Harness added value is to integrate these services into a unified DevOps platform, though others such as GitLab and GitHub are also on this path.