Expose wip

[bcomnes]

Sorry to spam so many PRs!

I ran into a bit of a footgun with the WIP event. I understood it to emit every 2 seconds when there was work de-queued into a worker, but apparently its only every 2 seconds max, between job transitions (which is normally way more often than 2s in a healthy queue!).

But if you have a large queue of work that takes 20 minutes each job, and you de-queue 15 jobs, and they all start to run for 20 minutes at the same time, you only really get WIP ~20 minutes.

Since I want to use WIP as an observability hook, only getting metrics every 20 minutes won't work for that use case.

This PR exposes the WIP method so that I can call it from the outside at whatever frequency makes sense for the metrics collector (in my case).

[Copilot]

Pull request overview

Exposes the internal “work-in-progress” (WIP) worker snapshot as a public API so external systems can poll worker utilization on-demand (instead of relying solely on the wip event cadence).

Changes:

• Add PgBoss#getWipData() that forwards to the manager’s WIP snapshot.
• Document getWipData() and recommend it for metrics/monitoring use cases with long-running jobs.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
src/index.ts	Adds a new public getWipData() method returning WipData[].
docs/api/workers.md	Updates worker monitoring docs and adds a new getWipData() API section with an example.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[coveralls]

coverage: 100.0%. remained the same — bcomnes:expose-wip into timgit:master