Development platform: GitLab 17.10 focuses on code review with AI

The March update for the GitLab development platform has been released. Version 17.10 extends the functions of GitLab Duo – for AI-supported code reviews and root cause analyses – and raises GitLab query language views to beta. A total of over 120 new features have been incorporated into the release.

GitLab Query Language Views enters beta phase

GitLab 17.10 introduces the beta version of GitLab Query Language (GLQL) Views for users of all GitLab editions. The GLQL views embed live data queries in markdown blocks in wiki pages, Epic descriptions, issue comments and merge requests. This should help teams to see the information they need and improve their collaboration without having to leave their current workflow.

Further information on the GLQL views can be found in the documentation.

Updates for the GitLab Duo AI service

One of the features of the GitLab Duo AI service is Duo Code Review. This has now reached beta status with GitLab 17.10. Developers can have code reviews based on artificial intelligence performed by adding @GitLabDuo as a reviewer to a merge request. By mentioning @GitLabDuo in comments, follow-up questions can then be asked.

Duo Code Review is available for GitLab Ultimate and GitLab Duo Enterprise customers and uses Anthropic Claude 3.5 Sonnet as the Large Language Model (LLM).

The development team has also worked on the GitLab Duo Root Cause Analysis: This feature is now available for Ultimate and Duo Enterprise customers of self-managed GitLab instances and is also in beta status. It is used to troubleshoot failed jobs in CI/CD pipelines: Root Cause Analysis is designed to analyze the log of the failed job, identify the cause and suggest a fix. According to GitLab, data sovereignty is not compromised in the process.

All further details on the new release can be found in the GitLab blog.

(mai)