Introduces a new SegmentIntrospection API endpoint and integrated tests to fetch metadata for stat segments. Tests cover multiple segments (release_count, ai_cost_cents, commit_count, diff_delta, etc.), validating properties like name, description, higher_is_better, unit, category, is_ratio, rounding_precision, and filters, plus available_aggregations including AI-related options when supported. This enables dynamic client UIs to discover capabilities and render segment data accurately.

To consider: Will this encumber the auto-focus of "PR awaiting review"?

Adds substantial_work? to CommitterChangelog and introduces a tl;dr Summary in the changelog UI. The changes, implemented in vendor/gems/chart_glimpses/app/models/committer_changelog.rb, vendor/gems/chart_glimpses/app/views/snap_changelogs/_change_logs.html.haml, and vendor/gems/chart_glimpses/app/assets/stylesheets/snap_changelogs/_change_logs.scss, add a tl;dr Summary header in the views and render a bold indicator for entries where substantial_work? is true; a new CSS block styles the summ...

This change updates ExternalUsageClient::ExternalCursorAiUsageClient to align with the updated Cursor daily-usage-data API. It now derives committers from either userId or email, aggregates usage into per-date LlmUsageDate records, and creates new records when identifiers cannot be mapped. It also introduces a one-time admin notification if a Cursor entry lacks an identifiable user, enabling proactive alerts for future API changes. Improves data accuracy and reliability of usage reporting.

This change enables editing of CAB changelogs and refactors how they are rendered and fetched. In recent-branches-prs-viewer.js, renderChangelogs now maps each changelog to either a ChangelogDetail or a ChangelogEditor and uses initialAfterMediaId. CommitsRecentActivity is updated to scope, preload, and limit changelogs (RECENT_CHANGELOGS_TO_SHOW_CAB) and to resolve media URLs with rails_public_blob_url for media_after/before. ChangelogEditor save messaging is simplified, and a minor CSS clea...

Adds a new API reference documentation page for surveys (polls) under the customer_api reference. The newly added vendor/gems/customer_api/app/views/api_reference/_surveys.html.haml documents the Surveys endpoints for creating surveys, managing questions, submitting and retrieving responses, and retrieving results. It includes parameter lists and example requests/responses to aid integrators. Benefit: helps developers discover and correctly use the Surveys API and reduces integration time.

Introduces grouping of AI usage tokens by provider in the AI usage settings UI. The server-side template at app/views/shared/settings/ai_usage.html.haml was updated to render tokens by provider, and the frontend styles in app/javascript/apps/ai-usage-token-settings.scss were extended to support the new layout. This change improves readability of token usage data and enables quicker cross-provider comparison for better cost and usage insights.

Fixes a UI bug where issue_browse displayed the 'All Time' entry twice. The patch updates the rendering in app/views/shared/resource_actions/issue_browse.html.haml and adjusts test/integration/issue_browse_test.rb to reflect the correct order: show 'All Time' once, followed by 'Per Interval' items. This eliminates duplication, clarifies the results for users, and improves test coverage.

Adds a deterministic ascending sort to AI Cohort charts by enforcing the order from Least use to Most use in Concerns::AiSegmentConcerns::derive_series_segments. The change, implemented in app/lib/presenters/concerns/ai_segment_concerns.rb, adjusts the series derivation to produce a consistently ordered, readable chart that better communicates usage distribution across cohorts.

This change updates TeamRepo::ensure_cached_stat_rebuild_on_create_destroy to queue an LLM metric build when a new team is created that may have LllUsageMetrics. It adds logic to trigger a cached stat rebuild for relevant teams without altering existing behavior. The result is earlier preparation of LLM usage metrics, improving visibility and responsiveness for teams likely to require these metrics.