Files
Zach e1c9525f4e Use T-digest for health latency monitoring (#10897)
## What changed?
Added a time-windowed tdigest implementation to allow for local
collection of health-related statistics. These will be used to report on
system health via the DeepHealthCheck endpoints on the history and
frontend services.

## Why?
The existing moving averages are not sufficient to track cluster health
accurately. T-digest allows for the collection and measurement of
trimmed-mean and quantile measures.

## How did you test it?
- [x] built
- [x] run locally and tested manually
- [ ] covered by existing tests
- [x] added new unit test(s)
- [ ] added new functional test(s)

---

Note: Branched from https://github.com/temporalio/temporal/pull/9940

---------

Co-authored-by: Nick Beaumont <nicholas.beaumont@temporal.io>
2026-07-10 10:31:14 -07:00
..