[Meta] Health Report - future phases
- Dominant language
- Java
- Stars
- 14.9k
- Forks
- 3.5k
- Avg merge
- 19h 14m
- Merged PRs (30d)
- 63
Description
Following the release of the #16056 in Logstash `8.15`, and subsequent addition to the Logstash integration, the first phase of the Logstash Health Report was completed.
This meta issue will capture additional work for future iterations of the Logstash health report 👍
### Additional pipeline probes
- [ ] periodically populate indicator details with immutable snapshot of pipeline flow metrics
- [ ] introduce expression-based probe type with configs `degraded`,`critical`, and `recovery`
- [ ] add `queue_persisted_growth_events` probe
- [ ] add `queue_backpressure` probe (requires wall-clock reorientation of underlying flow metric)
- [ ] add `input_throughput` probe
- [ ] add `output_throughput` probe
### Resources Indicator
We will provide a top-level indicator to highlight issues with the process's use of resources as a whole.
- [ ] create `resources` indicator
- [ ] create probe `resources:gc_pause_ratio` based on a new flow metric
### Iterative Improvements
We will iterate on the API to improve its usefulness, starting with:
- [ ] add support to `?verbose=false` to `GET /_health_report` to avoid populating indicator `details` when building the report
- [ ] add support for `GET /_health_report/pipelines/`
- [ ] add support for not propagating an indicator's degraded status (settings)
- [ ] add support for ACK-ing a currently degraded indicator status (API)
- [ ] add report reference, to include along-side anywhere we report status, and allow fetching a specific report for a reasonable ttl.
Contributor guide
Assessment
This issue has not been assessed yet.