Platform
The queue is the bottleneck. Here is who is in it.
Delivery rarely slows down because people write code more slowly. It slows down because pull requests sit unopened, because two people are the review queue for everybody, and because nothing surfaces the four PRs that have been blocking a release for a fortnight. Pacia surfaces all three.
- Waiting for review
- 17 open
- Stale over 24 hours
- 9
- Aged over 30 days
- 3
- Median pickup time
- 9.8h
- Top reviewer share
- 41% of all reviews
What you can see
The stuck-work queues
Filtered views for all, our team, participating, stale beyond 24 hours, aged beyond 30 days, waiting for review, and bots. The list you would otherwise rebuild by hand each Monday.
Pickup time
The gap between a pull request opening and the first review starting. Usually the largest and least-examined stage in the whole cycle.
Reviewer concentration
How much of the review load sits with how few people — the measure that predicts a burnout conversation months in advance.
Pull request size and rework
Size distribution and how often a PR gets substantially reworked after review. Large PRs are slow to review and more likely to fail; this is where the evidence for a size agreement comes from.
Merge quality
A dedicated view of merge outcomes with its own explore drill-down, for teams tightening up what is allowed to land.
Per-PR detail
Any pull request opens to a full timeline — reviews, comments, checks, stage timings — so a number never has to be taken on trust.
Start with the signal. Open the work behind it.
PR Insights shows where flow is slowing, then lets you move from the organisation view to the exact pull request, review, check or deployment event that produced the number.
How PR Insights moves from a pattern to its evidence
Example dataSee the pattern
- Waiting for review
- 8
- Median pickup
- 7.4h
- Stale over 24h
- 3
- Top reviewer share
- 38%
Pickup time is above the team target
Choose what to inspect
- Waiting queue8 pull requests→
- Cycle stagesCoding · pickup · review · merge→
- Reviewer loadWho the queue depends on→
- Merge qualityChecks, rework and outcomes→
Open the evidence
- PR openedMon 09:14
- First reviewTue 03:26 · 18h later
- 14 checks passedTue 06:39
- Merged & deployedTue 14:08
The 7.4-hour median is never a dead-end score. Open the queue, select a pull request and see the reviews, comments, checks and deployments that made up its timeline.
How the data arrives
Real time on the GitHub App
Repositories connected through the Pacia GitHub App receive signed webhooks for pull requests, reviews, comments and pushes. They are verified, queued and processed asynchronously, so the endpoint responds instantly and nothing is lost on a restart.
Every 20 minutes on tokens
GitLab, Azure DevOps, Bitbucket and GitHub-via-token repositories are reconciled on a 20-minute cycle that re-pulls the most recently updated pull requests. There is also a manual refresh button per repository.
180 days backfilled at connection
A background job pages through up to 180 days of history the moment a repository is connected, so you have trends immediately rather than in three months.
One shape across four providers
Each provider has an adapter that maps its own API into a single normalised pull request model before anything is stored. A cross-provider comparison is a real comparison.
Common questions
Does this work across providers at once?
Yes. An organisation can have GitHub, GitLab, Azure DevOps and Bitbucket repositories connected simultaneously, and teams can span them. Every provider is normalised into the same pull request shape before storage.
Do bots pollute the numbers?
Bot pull requests are identifiable and filterable rather than silently included. Bot and service accounts also never count towards billing.
Can I get the stuck queue in Slack?
Yes. The Slack and Microsoft Teams apps post the pull request queue, lifecycle alerts and a configurable pickup warning into a team’s channel. Many teams work entirely from the channel.