Adds default-branch validation for commits in closed PRs and adds tests. Updates to app/jobs/process_closed_pull_request.rb and app/models/commit_site.rb (kin_on_default_branch?) enforce default-branch checks when determining merged commits, with tests in test/jobs/process_closed_pull_request_test.rb. Improves PR processing reliability by preventing merges from non-default branches.

Adds provider_email to the developer_list data and introduces a fallback lookup by email when the provided git_provider_id does not match any known developer. This change updates the API response logic so provider_email is returned consistently and prevents failures when IDs are out of sync. Affected areas include CustomerApiBaseController (response assembly) and ApiReferenceHelper::developer_list_response to expose the new field, enabling more reliable client data downstream.

Adds a new stat: teammates_review_minutes, wired into LlmTeamCohortStat via a new committer_stat segment to quantify how much teammates spend on reviews when AI is involved. Validation for start_date and date_type_em is extracted from StatCacheConcerns into a standalone concern since AI stat models don't rely on digests. Main file updated: vendor/gems/llm_metrics/app/models/concerns/llm_cohort_stat_calculation.rb. This enables measuring AI-driven review burden and improves workload visibility.

Introduces a new unit test github_copilot_different_entity_test.rb to validate processing of Copilot stats across Bitbucket, Gitlab, and Azure organizations. The change also updates app/models/provider_resource_mirror.rb to align with the new processing flow exercised by the test. This boosts test coverage and reliability of Copilot metrics handling across providers, with the benefit of catching integration issues earlier and supporting smoother deployment of production features.

Added a test to verify that Copilot stats can be processed for Bitbucket, Gitlab, and Azure organizations. The test exercises the processing path end-to-end and passes in CI, giving greater confidence against regressions. No specific application files are indicated as changed, so the update appears to focus on test coverage. Impact: improves reliability of Copilot stats processing across these providers and reduces risk of regression in production.

This change refactors the ImportRepo job to improve reliability and throughput: it now makes a final import attempt when a valid git connection is found, applies exponential backoff with jitter to reduce quota clashes under large queues, and augments customer_error_detail with a clearer message when an import limit is reached. This results in more resilient imports under peak load and better visibility for customers.

In app/models/repo_import.rb, RepoImport.to_api_json now attaches detailed error information to the List Import API response whenever available. This provides clients with clearer failure reasons instead of generic messages, improving debuggability and speeding issue resolution in production.

In app/jobs/import_repo.rb, the ImportRepo::perform flow is refactored to perform a final import attempt when a valid git connection can be established, add exponential backoff with jitter to soften bursts when large import queues are present, and append a clearer message to customer_error_detail when an import limit is reached. This improves reliability for high-traffic scenarios and provides better fault visibility and user guidance.

Revise AI stats so they automatically aggregate by the method that yields the most interesting results, with an eye toward extending the model to ResourceTeamStat. Update tests in ai_impact_charts_test.rb to reflect variation in the optimal aggregate type and resource, and alphabetize StatCacheConcerns class and instance methods. Retention_window_start_date was moved into stat_cache_concerns.rb. Benefits: more flexible analytics, better test coverage, and groundwork for broader stat modeling.

Revamps AI stats to automatically select the aggregation method that yields the most interesting results, with a planned extension to ResourceTeamStat. Updates ai_impact_charts_test.rb to reflect variation in the optimal aggregate type and resource. Alphabetizes class and instance methods of StatCacheConcerns to improve consistency and maintainability. Benefits include more expressive AI metrics and smoother future expansion.

Introduces the 2025 developer productivity research on the Research page by updating app/views/pages/snippets/_first_party_research.html.haml and adding the new image asset app/assets/images/pages/ai-developer-productivity-change-2025.png. The changes render the study content within the Research page, providing current productivity insights to guide decisions and enhance developer resources.

Introduces a new 2025 developer productivity research entry on the Research page as a feature addition. Authored by Bill Harding and released to production on Oct 10, 2025. This update enhances the Research page with current productivity insights, improving access to relevant research for developers and stakeholders.

Introduces a new Stats::AiHistory (app/lib/stats/ai_history.rb) implementing the Stat Calculator paradigm to drive the feature-chart JSON boxes. Adds four initial metrics—All suggestion, Tab acceptance %, Prompted acceptance %, and Engaged user count—along with wiring for Engaged user count and Inactive AI user count in the analytics flow. Also adds Defines::Interval.derive_optimal_interval to derive dates with the optimal interval. Enables tracking engagement and inactivity in charts for bet...

Refactors GenerateResourceTeamStat (vendor/gems/stat_master/app/jobs/generate_resource_team_stat.rb): replaces ensure_updated_committer_counts with ensure_team_committer_activity and calls it when committer segments are requested; backfills TeamCommitterActiveDate, verifies timezone offsets, tightens in-repo committer filtering, and scopes expunge_outdated_digests to the requested segments. Result: fewer stale digests and more accurate per-team metrics.

Fix in ExternalClient::GitlabClient#user (vendor/gems/repo_interface/app/lib/external_client/gitlab_client.rb) to reliably fetch a GitLab user by ID. The changes refactor and reorganize the user lookup logic, adding the necessary path to locate a user by Gitlab ID and aligning behavior with existing API calls. Improves reliability and correctness of user resolution by ID in production.

Fix Starred Reports min data segment not being applied to single-segment charts. This change updates app/lib/presenters/concerns/cached_stat_graph_value_calculate_concern.rb, refining as_json and all_ordered_segments handling to ensure the minimum data constraint is enforced for single-segment charts as well as multi-segment ones. This resolves the test failure and ensures accurate rendering and consistent chart data.

Reintroduces the committer activity indicator in CAB by moving the styling from cab-committers-frame.scss into a new committer-activity-icon.scss and applying related updates. Main files touched: cab-committers-frame.scss and the new committer-activity-icon.scss. This restores visibility of who committed changes in the CAB viewer, improving clarity and traceability without altering overall UI behavior.