Enforce sorting in derive_series_segments (app/lib/presenters/concerns/ai_segment_concerns.rb) so cohort series are ordered from 'Least use' to 'Most use'. Updated series derivation and ordering logic to eliminate inconsistent ordering; charts now consistently display increasing usage, improving readability and correctness of AI cohort visualizations.

Update TeamRepo.ensure_cached_stat_rebuild_on_create_destroy (gitclear_teams) to enqueue an LLM metric build when a team is created that might include LLM usage metrics. Ensures LLM usage stats for new teams are generated promptly and available for reporting and monitoring.

Return audit PRs using the PR's most recent merged_at when a closed-date filter is provided, and expose new audit attributes: business_minutes_awaited_first_review, business_minutes_cycle_time, business_minutes_lead_time, business_minutes_open_to_closed, first_commented_at, first_commit_authored_at, post_merge_delta, and squash_commit_sha. Changes touch V1::AuditsController and pull_request_json_concerns to surface these values. Benefit: more accurate closed-date filtering and richer timing/p...

Updated app/jobs/assign_pull_request_post_merge_commit_work.rb (assign_post_merge_by_pull_request) to broaden detection of commit-message lines that indicate a squashed PR was updated and adjusted matching/handling so squashed-PR post-merge commits are recognized. Corresponding unit test updated in test/unit/pull_request_post_merge_commit_test.rb. Resolves failing test and ensures correct association of squashed PR commits after merge.

In app/models/repo_import_progress.rb, an early exit is added to calculate_percent_complete to short-circuit the evaluation when the default branch has already progressed to the repository's processing epoch. This reduces unnecessary work, improves performance, and ensures the progress percentage reflects epoch-boundary conditions during repository imports.

Introduce an early return in app/models/repo_import_progress.rb::calculate_percent_complete to short-circuit evaluation when the default branch has been processed back to the repo's provider_created_at. This skips redundant computation and prevents overcounting work, improving import progress accuracy and performance.

This change refactors how time-off data is built for the CAB UI: the derive_time_off_array logic, previously in Ruby (vendor/gems/commit_activity_browser/app/lib/bubble_committer_concern.rb), has been moved into the JavaScript client (app/javascript/bundles/cab-viewer/cab-committers-frame.js) and renamed the payload from timeOffHtml to timeOffArray. The move reduces HTML escaping issues in CAB by producing a data array on the client side, and updates the committer tooltip rendering to consume...

Fix bug in fallback calculation that failed to divide 30.seconds by 1.minute, which produced implausible minute estimates that propagated to PullRequestReviewer. Updated derive_minutes_used_details in vendor/gems/comment_reactions/app/models/code_comment_extra.rb and added/adjusted tests in test/unit/pull_request_comment_test.rb to validate the corrected behavior. Result: realistic comment-time estimates and accurate reviewer metrics.

This feature adds a sparkline visualization to the CAB committer tooltip, showing the delta of activity over the past year when hovering a committer avatar. Implemented in cab-viewer/cab-committers-frame.js by extending renderCommitter::committerTooltipHtml::renderSparkline and related tooltip content, enabling a compact visual summary of recent contributions and improving quick context for reviewer or teammate.

This change increases how often PR discovery is queued in the early life of a repo, since the DiscoverPr job no longer scans beyond the main branch plus a buffer. It also refactors and reorganizes several subscription methods (notably in app/lib/defines/subscription.rb and vendor/gems/subscribe_manager/app/models/subscription.rb) to clarify behavior and improve refresh timing. Result: faster, more reliable PR discovery for new repos and more predictable subscription timing.

Adds automatic queuing of PR discovery when creating a release rule that targets pull requests. Implemented in vendor/gems/release_track/app/models/repo_release_rule.rb (after_create_rule and related PR-rule helpers). This ensures PRs are discovered immediately upon rule creation, reducing latency and improving reliability of PR-based release rules.

Adds a new aggregate_by parameter to the Reports API, enabling clients to specify how report results are grouped. The API surface is extended in vendor/gems/customer_api/app/lib/defines/api.rb and the v1 reports controller in vendor/gems/customer_api/app/controllers/api/v1/reports_controller.rb to accept and propagate the parameter. This change provides more flexible, client-driven aggregation without altering data retrieval, improving reporting versatility for dashboards and integrations.

Introduces a new API endpoint to submit AI usage statistics (POST /api/v1/ai_usage) and adds a corresponding integration test. The test file vendor/gems/customer_api/test/integration/api_ai_usage_test.rb exercises the endpoint to verify request handling, payload validation, and responses. This improves observability of AI usage and provides automated validation in CI; production release occurred on Nov 11, 2025.

Adds periodic reprocessing for releases and hooks to refresh defects and issues. Reorganizes release population logic: most code moved from the previous rule to vendor/gems/release_track/app/jobs/populate_releases_from_rule.rb, with perform/queue paths consolidated. Introduces a new vendor/gems/release_track/lib/tasks/releases.rake to drive periodic processing. For active subscriptions, refresh frequency increases and ResourceTeamStats updates are limited to subscribed entities, delivering fa...

Adds a new job (app/jobs/commit_default_branch_confirmer.rb) that introduces a redundant call site to mark commits as present in the main branch when the commit site is kin in the default branch. Some perform logic from CommitCompleteJob is now moved into the new file and commit_site wiring is updated accordingly. This delivers more robust, consistent labeling of commits relative to main versus default branches, improving downstream processing.

Adds two AI Cohort stats items, Test Delta Cohorts and Copy/Paste Cohorts, to the AI Cohort Stats lineup. Updates the chart color palette with new palettes (cyberPunk, doomPalette, fuchsiaPalette, etc.) and fixes color retrieval in AreaChart by deriving the color set when the segment parameter isn’t passed. Main changes are in app/javascript/bundles/area-chart/area-chart-segment-colors.js; UI reference updated in ai_cohorts_stats.html.haml. Impact: clearer visuals and more reliable theming fo...

Aligns the AI Cohorts statistics UI strings by replacing the legacy label 'Commit count per AI Cohort' with 'Commit Count per Count in AI Cohort' across the UI and its presenters. Updated occurrences in app/lib/defines/stat_segments.rb, app/views/shared/resource_actions/ai_cohorts_stats.html.haml, app/lib/presenters/historic_ai_cohorts_stat.rb, and app/lib/helper_concerns/application_resource_helper_concerns.rb to ensure consistent capitalization and wording, improving clarity and reducing us...

Rework the redirect decision logic in the repo import flow so that, when there is no evidence that a user has completed the GK redirect, the GK redirect path takes precedence over other redirect types (except the usage/profile path). This change touches app/lib/controller_concerns/repo_import_concern.rb and app/lib/git_kraken.rb, adjusting the priority rules and how completion evidence is evaluated. Benefit: more predictable, correct redirects when GK completion is uncertain, reducing incorre...

Consolidates the GitKraken redirect logic from UserUsageProfilesController into a new app/lib/controller_concerns/kraken_controller_concern.rb as KrakenControllerConcern. The moved methods, including notify_gitkraken_and_redirect and confirm_provider_import_with_gitkraken, are now accessible via import or the user usage profile, improving separation of concerns and testability. This modularization enhances reuse and maintainability of the GK redirect behavior.