The GitHub Actions job "Build documentation" on flink-kubernetes-operator.git 
has succeeded.
Run started by GitHub user asfgit (triggered by asfgit).

Head commit for run:
ad8b4abf7334c7f196bd5a43f43e6181c9466574 / Maximilian Michels <[email protected]>
[FLINK-31326] Consolidate source scaling logic (#546)

Previous changes removed the requirement for sources to have the pendingRecords 
metric. This allows to refactor and consolidate the source scaling code. We 
remove the option to disable scaling sources because the primary motivation was 
to skip scaling legacy sources which we can achieve without an extra 
configuration option.

The refactored source scaling logic works like this:

* Modern sources (busy time available)

Scale source based on the busy time up. Add extra capacity if the lag
information, provided by the pending records metric, is available. Otherwise
assume 0 lag. Cap by the maximum number of partitions, if available.

* Legacy sources (busy time unavailable)

Leave the source parallelism unchanged, use the current processing rate as the
target rate downstream.

Report URL: 
https://github.com/apache/flink-kubernetes-operator/actions/runs/4400082090

With regards,
GitHub Actions via GitBox

Reply via email to