Problem Definition

Currently, Bitbucket Data Center lacks a centralized method for administrators to gather detailed feature usage statistics at the service level. Administrators can't easily access metrics such as the number of projects or repositories with specific features enabled (for example, webhooks, automated merging, build before merging), user interactions with features (for example, code search during PR review, online code editing), or configuration preferences (for example, signed commits, group vs. user level access). This information is critical for understanding feature adoption, optimizing resource allocation, and enhancing user productivity.

Suggested Solution

Implement a set of Admin APIs within Bitbucket Data Center that provide comprehensive usage statistics and configuration details at the service level. These APIs should allow administrators to:

• Retrieve counts and details of projects/repositories with specific features enabled.
• Access user interaction metrics, such as usage of code search or online editing.
• Obtain configuration statistics, such as the number of repos with signed commits or group vs. user-level access.
• Aggregate data across teams, projects, and repositories for a holistic view of feature usage.

Why This Is Important

• Access to detailed feature usage statistics empowers administrators to:
• Make informed decisions about resource allocation and feature prioritization.
• Identify underutilized features and assess training needs or potential feature deprecation.
• Monitor compliance with organizational policies, such as commit signing or access controls.
• Enhance overall productivity by aligning feature development with actual user needs and usage patterns.

Workaround

Currently, administrators can obtain some of the desired statistics through:

• Manually querying the Bitbucket database for specific configuration details and usage patterns.
• Using the available REST API endpoints to gather limited information on a per-repository or per-project basis.
• Parsing access log files to infer user interactions and feature usage, although this method can be labor-intensive and imprecise.