influxdata / influxdata/kapacitor
Recording returning immediately from aborted edge, no data and no error
- Dominant language
- Go
- Stars
- 2.4k
- Forks
- 479
- Avg merge
- 4d 16h
- Merged PRs (30d)
- 4
Description
The behavior is that running:
```
kapacitor record stream -task cpu_alert -duration 20m
```
returns immediately, the status of the recording is `finished` but logs show the edge was aborted and no error shows up in the logs either.
Task:
```
stream
// Select just the cpu measurement from our example database.
|from()
.database('cp')
.measurement('user_cp')
|window()
.period(15m)
.every(15m)
|alert()
//.crit(lambda: "usage_idle" < 100)
// Whenever we get an alert write it to a file.
.info(lambda: "cp" > 150.0)
.log('/tmp/alerts.log')
```
[Logs](https://github.com/influxdata/kapacitor/files/533960/CyQN1Eyf.txt)
[Stack Trace](https://github.com/influxdata/kapacitor/files/533961/DAqRwUbi.txt)
Contributor guide
Research direction
Reproduce the issue with `kapacitor record stream -task cpu_alert -duration 20m` and review the linked logs and stack trace first. Trace why the recording reports `finished` after the edge is aborted; it is done when the failure is surfaced as an error instead of silently returning with no data.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- cli, stream-processing
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100