influxdata / influxdata/kapacitor
[feature request] Improve handling of data gaps
- Dominant language
- Go
- Stars
- 2.4k
- Forks
- 479
- Avg merge
- 4d 16h
- Merged PRs (30d)
- 4
Description
At the moment, Kapacitor does not do well when points are not being received. For example, when using `|join()` against multiple input streams, the join node cannot recognize the absence of points (necessary for `.fill()`) until more points are received. This can be limiting in environments where data streams may be sparse (it could be minutes, hours, days before more points are received) and alerts need to be fired before the next points arrive. It would be great if Kapacitor had a more intelligent method for filling in data gaps, for example a `|fill()` node (or an improved `|default()`).
Contributor guide
Research direction
Start by tracing the existing |join() behavior and how .fill() and |default() handle missing points in sparse streams. Compare their current behavior with the alerting needs described here, then define and test an explicit data-gap policy before implementing a new or improved node.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- observability, stream-processing
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100