chia7712 opened a new pull request, #12528:
URL: https://github.com/apache/kafka/pull/12528
In our env, the throttled throughput is not stable in the beginning (see
below)
```
[2022-08-13T13:17:18.48906045]
Topic "bd5d82b62bd148e7887ddcc3361fc9f9":
| Partition 0:
| | replica on broker 1 => [ ] 0.00% 0.00
Byte/seconds (unknown) []
| | replica on broker 0 => [####################] 100.00% [leader,
synced]
[2022-08-13T13:17:19.488684125]
Topic "bd5d82b62bd148e7887ddcc3361fc9f9":
| Partition 0:
| | replica on broker 1 => [########## ] 50.69% 102.67
MB/seconds (about now) []
| | replica on broker 0 => [####################] 100.00% [leader,
synced]
[2022-08-13T13:17:20.488056226]
Topic "bd5d82b62bd148e7887ddcc3361fc9f9":
| Partition 0:
| | replica on broker 1 => [########## ] 53.25% 5.19
MB/seconds (18s estimated) []
| | replica on broker 0 => [####################] 100.00% [leader,
synced]
```
The root cause is `ReplicationQuotaManager` records throughput even though
there is no throttled partitions. Hence, the sample window created too early
leads to incorrect (smaller than actual) throughput.
### Committer Checklist (excluded from commit message)
- [ ] Verify design and implementation
- [ ] Verify test coverage and CI build status
- [ ] Verify documentation (including upgrade notes)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
