The following page may contain information related to upcoming products, features and functionality. It is important to note that the information presented is for informational purposes only, so please do not rely on the information for purchasing or planning purposes. Just like with all projects, the items mentioned on the page are subject to change or delay, and the development, release, and timing of any products, features or functionality remain at the sole discretion of GitLab Inc.
Last updated: 2023-11-06
Make it easy to build a feature into GitLab across multiple types of deployment.
Please reach out to Roger Woo, Senior Product Manager for the Cloud Connector group (Email) if you'd like to provide feedback or ask any questions related to this product category.
“Make it easy to build a feature into GitLab across multiple types of deployment”
GitLab Cloud Connector is a way to access services common to multiple GitLab deployments, instances, and cells. GitLab customers can choose between the following deployment options:
The goal of GitLab Cloud Connector is to ensure that GitLab features can work equally across all three. This aligns with Enablement’s mission of accelerating other product groups by reducing the friction of developing for multiple deployments.
Historically, GitLab has bundled new features directly into our self-managed application for customers to deploy as desired. However, newer capabilities require increased infrastructure requirements, making it difficult to offer in a self-contained environment. For example, AI features rely on specialized hardware provisioned through large cloud platform vendors. Database technologies such as Clickhouse are expensive and complicated to operate for smaller standalone instances.
Cloud Connector contributes to GitLab’s vision of an integrated DevSecOps platform by providing one single interface for customers to access multiple GitLab-hosted solutions, such as AI features, observability, and secrets management.
In our first iteration, GitLab Cloud Connector introduced the capability of “instance-level authentication” to GitLab. This allowed GitLab Inc to validate the identity of a self-managed instance and facilitate its access to cloud-based services. Our initial use case enabled Code Suggestions (beta) for self-managed and Dedicated customers.
As Cloud Connector evolves to support additional AI features, we have divided our team’s scope into three overarching domains: Cloud Connector Foundation contains the platform primitives shared across multiple features and products. AI-related Features encompasses our collaborative efforts to bring AI-related features to parity across SaaS, Dedicated, and self-managed instances. Future Opportunities represents Cloud Connector’s aspirational roadmap. These items are generally 6+ months in the future and often conceptual or speculative in nature.
It is important to note that GitLab Cloud Connector does not currently exist as a standalone technical system. We are working to formalize the technical boundaries of Cloud Connector by early 2024. A potential approach is to stand-up a dedicated Cloud Connector Service.
GitLab Cloud Connector is currently responsible for two foundational capabilities:
Our vision is to identify problems that span multiple feature groups and provide feature groups with a consistent out-of-the-box solution. We believe this will enable teams to focus on their domain expertise and accelerate GitLab’s innovation velocity.
For example, we are exploring solutions for application level rate limits and minimizing customers permit-list configurations. GitLab-hosted solutions will likely be backed by some common underlying resource and we believe GitLab Cloud Connector is well-positioned to mitigate a “tragedy of the commons” between instances. We can also streamline customer onboarding by providing a single entry point into multiple GitLab-hosted features.
In the long-run, we will work with other platform groups to support capabilities such as usage-based billing and regional routing for customers with data residency concerns.
Our immediate focus is to bring parity across SaaS, self-managed, and Dedicated for all AI-related features. AI-related features will generally be available on SaaS prior to self-managed / Dedicated.
We are prioritizing AI feature development on SaaS due to its experimental nature and the high iteration velocity of GitLab SaaS. This reduces the number of changes that will ship to self-managed instances, and reduces the overall frustration of upgrades and deprecations. We believe this will yield a better experience for our self-managed customers.
These items are exploratory and over 6 months away in the future. Candidate products may address themes of remote development, observability / analytics, natural language search, and native secrets management. We currently have low capacity to support additional Cloud Connector use cases.
The Cloud Connector team does not own any customer-facing DevSecOps features. Instead, we provide an opinionated interface for GitLab customers to access features developed by other groups. For example, the Cloud Connector team enabled access to Code Suggestions (beta) for self-managed and Dedicated customers. However, the underlying foundational AI components and Code Suggestion functionalities are owned by AI Framework and Code Creation teams.
Our unit of focus is at the level of GitLab instances (self-managed, Dedicated, or SaaS). As such, our target persona is the GitLab instance administrator. We are not focussed on individual end-user personas.