gyfora commented on PR #686:
URL:
https://github.com/apache/flink-kubernetes-operator/pull/686#issuecomment-1771263621
> If we catch up during the stabilization period, then there is no
overestimation of the true processing rate. I thought this change is to prevent
scenarios where we are are overestimating our true processing rate, hence never
catching up?
@mxm
Imagine the following scenario:
1. The TPR is overestimated, lag starts to build up
2. Once there is enough lag, observed TPR kicks in, we rescale
3. We already catch up during stable
4. Since there is no observed rate, we use busy time -> we overestimate
(which was the original issue)
5. We scale down due to overestimation -> Go to 1.
--
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: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
