Pull Request Explorer

Summary

Add OTLP metrics for replay and replay vision

Enables dual OTLP metric emission for session replay and replay vision, improving observability and billing insights.

Health Assessment

• • PR merged quickly with minimal rework, but large code changes across backend and nodejs, indicating significant scope but smooth review.

Summary

Fix malformed flag-dependency conditions to avoid network calls

Ensures malformed flag dependency conditions are treated as no match locally, eliminating unnecessary /flags network calls and reducing traffic for customers. Adds comprehensive tests to verify zero network traffic for such cases.

Health Assessment

• • Fast turnaround with minimal review, indicating a straightforward bug fix with clear tests.

Summary

Move autocapture opt-in to settings

Centralizes error tracking opt‑in into a dedicated model, reducing hot table writes and laying groundwork for future feature rollouts.

Health Assessment

• • Fast cycle time but required multiple commits to address backfill logic and locking, indicating moderate complexity. The refactor improves data model hygiene without affecting current behavior.

Summary

Update E2E tooltip selectors

This change updates tooltip selectors to match the new Quill chart portal attribute, fixing deterministic E2E failures. It ensures reliable end‑to‑end testing of chart tooltips.

Health Assessment

• • Fast cycle time and minimal rework indicate a straightforward bugfix with low risk.

Summary

Prefetch alert threshold and subscribers to reduce queries

Reduces database queries for dashboards with alerts, improving load performance by up to 72%.

Health Assessment

• • Quick fix with minimal review, demonstrating efficient N+1 query detection and regression testing.

Summary

Make all Web Analytics notifications archivable

Adds an archivable flag to Web Analytics notifications, enabling users to archive them and reduce clutter.

Health Assessment

• • Quick, single‑commit PR with immediate approval and no blockers; minimal code churn and low risk to production.

Summary

Fix React Native repeating surveys

Ensures surveys reappear after new iterations by aligning seen-state logic across SDKs, improving user engagement and consistency.

Health Assessment

• • PR required extensive refactoring and multiple iterations, indicating significant complexity; however, rapid initial review and small number of comments suggest focused feedback.

Summary

Update research mode model to Claude Opus 4.6

Updates documentation to reflect the new AI model version, ensuring users see accurate information about the research mode's underlying technology.

Health Assessment

• • Fast merge with minimal changes, no review needed, low risk.

Summary

Expose S3 file change errors on warehouse reads

Provides actionable error messages for S3 file race conditions, reducing support tickets and improving user debugging experience.

Health Assessment

• • Quick fix with minimal rework; improved error visibility reduces support overhead.

Summary

Generalize eval harness for multiple suite types

Enables cheaper one‑shot evaluations and product‑owned suite discovery, reducing infra costs and expanding AI testing flexibility.

Health Assessment

• • Rapid iteration with 10 commits and a single review round, achieving a 9‑hour cycle time. The large scope and many commits suggest careful testing, but the quick turnaround indicates an efficient review process.

Summary

Make eval engine swappable via neutral interfaces

Replaces braintrust‑specific types with engine‑neutral abstractions, enabling future evaluation engines to be swapped in without code changes and improving maintainability of the eval harness.

Health Assessment

• • The PR was merged within 36 minutes of opening, indicating minimal review friction. The refactor introduces a large codebase change (over 600 lines) but is well‑tested, reducing risk.

Summary

Add real-time signup enrichment workflow

Enables real-time firmographic enrichment after signup, improving onboarding routing without blocking signup.

Health Assessment

• • PR involved multiple AI-generated commits and extensive rework after review, leading to a slow cycle time and high risk.

Summary

Fix MCP analytics failures source mismatch

Aligns failure reporting with the $mcp_is_error flag, ensuring accurate failure counts and labels for MCP tools.

Health Assessment

• • The PR was reviewed quickly (0.1h) but required a substantial code change (276 lines across 13 files), indicating a significant bug fix that could impact multiple components. The large scope combined with a single review round suggests a high risk of undiscovered side effects, despite the fast review.

Summary

Implement Fly.io data warehouse source

Adds a Fly.io source to sync apps, machines, and volumes into PostHog’s warehouse, enabling infrastructure inventory queries alongside product data.

Health Assessment

• • PR had a quick review and single commit, but large code addition; moderate risk due to size but low review friction.

Summary

Implement DigitalOcean data warehouse source

Adds full DigitalOcean integration for infrastructure and billing data, enabling users to sync cloud resources into PostHog Data Warehouse.

Health Assessment

• • Large code addition with a long cycle time but minimal review iterations, indicating a complex feature added with limited review feedback.

Summary

Fix free-cap for seatless posthog_code users

Ensures seatless users are capped at the $75 free plan limit, preventing them from exceeding default usage caps and aligning billing with intended free-tier behavior. This protects revenue and maintains system stability.

Health Assessment

• • Rapid 1.2‑hour cycle with multiple commits after review shows active rework but quick review turnaround; medium scope and backend changes imply moderate risk.

Summary

Add AI observability aliases to system prompt

Adds alias mapping for AIO, LLM analytics, and LLMA to improve agent terminology resolution, reducing confusion and enhancing user experience.

Health Assessment

• • Quick, one-line change with minimal review, AI-assisted writing, low risk and fast cycle time.

Summary

Add MCP tool call usage reporting

Enables billing to report MCP tool call events separately, improving adoption visibility for organizations.

Health Assessment

• • The PR was reviewed quickly (3.3h) and merged within 33h, indicating a smooth process. AI assistance was used to generate code and tests, reducing manual effort. The change is moderate in size (≈480 lines across 5 files) and introduces a new reporting metric without affecting existing functionality.

Summary

Add staff filter to view all team tasks

Enables staff to view any team task, bypassing per‑user visibility, improving support and debugging.

Health Assessment

• • Fast cycle time (2.5h) and minimal review friction indicate a straightforward, low‑risk feature addition.

Summary

Fix partition key handling in data warehouse

Corrects repartition failures caused by mismatched partition keys, improving reliability of data ingestion.

Health Assessment

• • Quick, single-commit fix with minimal code changes and no new features, resulting in a fast review cycle and low risk.

Summary

Fix missing incremental field error classification

Prevents retry loops and reduces error‑tracking noise by classifying a deterministic configuration error as non‑retryable, giving customers clear guidance to resolve the issue.

Health Assessment

• • Fast turnaround with a single commit and no review back‑and‑forth, indicating a straightforward, low‑risk fix.

Summary

Fix Anthropic retry-after handling for rate limits

Ensures data imports respect Anthropic's Retry-After header, preventing premature retries and reducing failures.

Health Assessment

• • Quick fix with minimal rework, approved in under 30 minutes, indicating low complexity and high confidence.

Summary

Shrink Anthropic usage report page size

Reduces page size to avoid 400 errors, ensuring stable data syncs.

Health Assessment

• • Quick, single‑commit fix approved immediately with no rework, indicating low complexity and risk.

Summary

Anchor Jest dev-folder ignore pattern to repo root

Fixes test discovery when the repository is checked out under a dev directory, ensuring Jest correctly runs tests and ignores only in‑repo dev folders.

Health Assessment

• • Quick fix with minimal changes, fast review, no rework.

Summary

Merch launch video

Adds a launch video to the merch page, automatically opening the first collection item to showcase new merch and boost user engagement.

Health Assessment

• • Fast cycle time of 22 hours, minimal rework, and small scope indicate a smooth, low-risk feature addition.

Summary

Remove third-party apt sources from CI workflows

Prevents CI failures caused by expired third-party APT repositories on Depot runners, ensuring reliable package updates.

Health Assessment

• • Quick fix to CI pipeline, minimal changes, resolved in under 3 hours, low risk.

Summary

Improve query schema validation with discriminated unions

Ensures API payloads for insights queries are validated against correct union types, reducing error noise and improving developer experience. This change provides clearer error messages for malformed requests, enhancing reliability for API consumers.

Health Assessment

• • R
• • a
• • p
• • i
• • d
• •
• • i
• • t
• • e
• • r
• • a
• • t
• • i
• • o
• • n
• •
• • w
• • i
• • t
• • h
• •
• • m
• • i
• • n
• • i
• • m
• • a
• • l
• •
• • r
• • e
• • v
• • i
• • e
• • w
• •
• • c
• • y
• • c
• • l
• • e
• • s
• •
• • i
• • n
• • d
• • i
• • c
• • a
• • t
• • e
• • s
• •
• • c
• • o
• • n
• • f
• • i
• • d
• • e
• • n
• • c
• • e
• •
• • i
• • n
• •
• • c
• • h
• • a
• • n
• • g
• • e
• • s
• • ,
• •
• • b
• • u
• • t
• •
• • t
• • h
• • e
• •
• • l
• • a
• • r
• • g
• • e
• •
• • c
• • o
• • d
• • e
• • b
• • a
• • s
• • e
• •
• • m
• • o
• • d
• • i
• • f
• • i
• • c
• • a
• • t
• • i
• • o
• • n
• • s
• •
• • a
• • n
• • d
• •
• • A
• • P
• • I
• •
• • v
• • a
• • l
• • i
• • d
• • a
• • t
• • i
• • o
• • n
• •
• • s
• • h
• • i
• • f
• • t
• •
• • p
• • o
• • s
• • e
• •
• • s
• • i
• • g
• • n
• • i
• • f
• • i
• • c
• • a
• • n
• • t
• •
• • r
• • i
• • s
• • k
• •
• • t
• • o
• •
• • d
• • o
• • w
• • n
• • s
• • t
• • r
• • e
• • a
• • m
• •
• • c
• • o
• • n
• • s
• • u
• • m
• • e
• • r
• • s
• • .

Summary

Shrink wizard payload and enable gzip

Reduces wizard endpoint payload by 93%, improving load times and cutting bandwidth usage.

Health Assessment

• • Quick, single-commit fix with minimal code changes and no new features, resulting in fast review and merge.

Summary

Add experimental date picker for insights

Introduces a chip‑based date filter UI for insights, enabling A/B testing of a compact design without backend changes.

Health Assessment

• • Rapid iteration with AI assistance; multiple commits after review indicate active refinement; overall fast cycle time.

Summary

Remove obsolete tooltip flag and legacy tooltips

Consolidates tooltip rendering to the new quill system, eliminating dead code and simplifying maintenance.

Health Assessment

• • Fast cycle time (4.5h) and minimal review indicate a straightforward maintenance task; the large number of deletions reflects a significant cleanup but with no new feature risk.

Summary

Add Data Ops Overview Tab

Provides users with a consolidated view of warehouse status, improving visibility into data backfill progress and readiness.

Health Assessment

• • Quick turnaround with a single commit and minimal review rounds indicates high confidence in the changes.

Summary

Surface billing activity in activity log UI

Adds billing activity scope and describer so billing events render meaningfully in the activity log, improving visibility for users.

Health Assessment

• • Quick turnaround with minimal rework; AI-assisted code generation streamlined implementation.

Summary

Render full summary as Slack section blocks

Ensures complete replay-vision reports are delivered to Slack without truncation, improving visibility and preventing data loss for users.

Health Assessment

• • Fast cycle time and single review round indicate efficient collaboration and minimal rework.

Summary

Add no-session-join fast paths for web analytics

Improves web analytics query performance by eliminating unnecessary session joins, delivering up to 15x speedups for key dashboards.

Health Assessment

• • Fast cycle time and minimal review rounds indicate efficient collaboration; performance gains are significant.

Summary

Release 166 fully implemented warehouse sources

Makes 166 data warehouse connectors visible to users by removing the hidden flag, enabling them to connect and use new data sources.

Health Assessment

• • Rapid approval with a single review and minimal changes; the PR was merged within 0.9 hours, indicating low complexity and high confidence.

Summary

Bump deltalake to 1.6.1 and pyarrow to 23.0.1

Upgrades deltalake and pyarrow to address type reconciliation issues in warehouse pipelines.

Health Assessment

• • The PR was thoroughly reviewed and tested, with a clear description of the changes and their impact on the warehouse pipelines.

Summary

Fix broken session replay link in blog post

Corrects a broken link in a blog post, ensuring users can navigate to the session replay page without encountering a 404 error.

Health Assessment

• • Quick fix with minimal changes, resolved within minutes, indicating efficient review and low complexity.

Summary

Add PostHog Code pricing page

Adds a pricing page for PostHog Code, detailing AI credit billing, free tier, model pricing, usage tracking, and controls.

Health Assessment

• • Rapid iteration with multiple small commits indicates active refinement; quick review and merge suggest low friction and high confidence in the documentation update.

Summary

Add threshold alerts for metrics insights

Enables threshold alerts on metrics insights by adding MetricsQuery as an alertable kind, updating backend extraction, validation, API gating, and frontend support.

Health Assessment

• • Rapid iteration with multiple commits and quick review indicates high developer engagement; large scope but short cycle time suggests efficient process.

Summary

Rename Django spans using route patterns

Improves trace readability by naming spans with resolved URL patterns, aiding debugging and monitoring.

Health Assessment

• • Fast review and merge with minimal changes; low risk and high confidence in correctness.

Summary

Preserve filename on uploaded media downloads

Ensures downloaded media retains original filename and extension, improving user experience and preventing file misidentification.

Health Assessment

• • Quick fix with minimal changes, fast review and merge, indicating low complexity and high confidence.

Summary

Add OAuth account picker for GitHub integration

Enables users to select GitHub repositories via a shared, searchable picker, improving UX and reducing custom code for the data warehouse integration.

Health Assessment

• • The PR was reviewed quickly and merged within 24 hours, but required a series of small commits to refine both backend and frontend logic, indicating a high iteration intensity but low overall scope.

Summary

Add person-target property editor for data warehouse

Enables creation of person properties from warehouse tables, adding UI for mapping and configuration, improving data integration.

Health Assessment

• • Fast review with a single commit, but the change spans over 500 lines across backend and frontend, indicating a substantial feature addition with moderate risk.

Summary

Add orchestration-agnostic enrichment core

Provides a reusable enrichment core for growth, enabling real‑time and batch orchestrators to enrich organization data with Harmonic and other providers, improving data quality and consistency.

Health Assessment

• • The PR required 4 days from first review to merge with a single comment and no rework, indicating a straightforward implementation but a relatively long cycle time likely due to scheduling or other dependencies.

Summary

Add ingestion control plane service

Provides an internal web tool to monitor ingestion lag and debug consumer routing, improving ops visibility and reducing manual troubleshooting.

Health Assessment

• • Rapid 4‑hour cycle with minimal review rounds; AI‑assisted code generation enabled quick delivery of a new internal ops service.

Summary

Update posthog-js dependency to 1.400.0

Bumps the PostHog JavaScript SDK to the latest 1.400.0 release, ensuring compatibility and access to new features.

Health Assessment

• • Quick dependency bump with minimal changes, fast review and merge.

Summary

Add unit tests for OAuth integration kind resolver

Adds comprehensive test coverage for the OAuth integration kind resolver, ensuring correct handling of nested OAuth fields and preventing 400 errors for unsupported sources.

Health Assessment

• • Fast review and merge, minimal changes, no production code affected.

Summary

Enforce Code Access for Cloud Runs

Prevents cloud task launches when the authenticated user lacks PostHog Code entitlement, ensuring secure access control.

Health Assessment

• • The PR was reviewed and merged quickly, indicating a smooth review process. The changes are well-documented and include regression coverage for the entitlement backstop.

Summary

Consolidate PR bot comments into shared CI report

Reduces noise in pull requests by merging separate CI bot comments into a single shared report, improving developer experience and reducing duplicate notifications. Adds a warning floor for dist folder growth to avoid false positives.

Health Assessment

• • The PR was merged quickly with minimal review, indicating efficient collaboration, but the large number of lines and files suggests significant refactoring of CI comment handling.

Summary

Add managed warehouse data status endpoint

Adds a new API endpoint that reports the readiness state of data in managed warehouses, allowing teams to monitor backfill progress and detect stalled partitions, improving visibility and operational confidence.

Health Assessment

• • The PR introduced a large backend feature with over 1,700 lines of code, but the review cycle was quick (2.6h to first review) and the changes were well-structured, indicating efficient collaboration. AI assistance was used to author key components, which may accelerate future iterations.