On February 20, 2025, GitLab released version 17.9, which introduced improvements aimed at enhancing user experience and functionality. A highlight of this release is the general availability of GitLab Duo Self-Hosted, enabling organizations to deploy large language models (LLMs) within their infrastructure.
This allows for the integration of GitLab Duo Code Suggestions and Chat using models hosted on-premises or in private clouds, supporting open-source Mistral models on vLLM or AWS Bedrock, Claude 3.5 Sonnet on AWS Bedrock, and OpenAI models on Azure OpenAI.
Organizations that deploy LLMs within their own infrastructure or private cloud environments avoid the risk of exposing proprietary code, intellectual property, or sensitive business data to external AI providers. Industries with strict compliance and regulatory requirements, such as finance, healthcare, and government sectors, benefit from this capability as they can leverage AI while maintaining full control over their data.
Joel Krooswyk, Federal CTO for GitLab, noted to DevOps.com, that although the software-as-a-service (SaaS) edition of the platform is seeing increased adoption, many organizations still opt for self-hosting due to factors such as regulatory requirements. He added that this self-hosted approach allows organizations to manage their own DevOps platforms and helps DevOps teams meet any data privacy requirements or concerns their organization may have.
Running LLMs on-premises or in a private cloud, organizations can eliminate latency associated with external API calls to AI services. This is especially beneficial for real-time AI applications. Additionally, regulatory concerns around data residency and compliance (such as GDPR, HIPAA, or SOC 2) are more manageable when AI processing remains within an organization’s controlled environment.
With AI-assisted coding and chat functions, organizations can embed secure coding practices into their development process. LLMs can help identify security vulnerabilities, suggest best practices, and even automate fixes before code is merged. This aligns with the growing trend of shift-left security, where security measures are integrated earlier in the development lifecycle.
Overall, GitLab 17.9 brings substantial improvements in AI integration, deployment efficiency, development environment collaboration, and project maintenance.
Another enhancement is the ability to run multiple GitLab Pages sites with parallel deployments, allowing for simultaneous updates to various sites, improving efficiency and reducing deployment times.
Integration capabilities have also been expanded with the option to add project files to Duo Chat within popular integrated development environments (IDEs) such as Visual Studio Code and JetBrains. This facilitates deeper code interpretation and collaboration directly within the development environment, which is likely aimed at enhancing productivity and teamwork.
To optimize project maintenance, GitLab 17.9 introduces the automatic deletion of older pipelines. This feature helps in managing storage and maintaining an organized project repository by removing outdated pipeline data, ensuring that resources are efficiently used.
AI continues to grow in support across the industry and this new release from GitLab showcases further progress in how development teams can leverage LLMs in new ways to enhance their efficiency.
%20Abstract%20Background%20102024%20SOURCE%20Amazon.jpg)
