You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "David Morávek (Jira)" <ji...@apache.org> on 2023/04/05 13:30:00 UTC

[jira] [Created] (FLINK-31736) Expose per-job capabilities via REST API

David Morávek created FLINK-31736:
-------------------------------------

             Summary: Expose per-job capabilities via REST API
                 Key: FLINK-31736
                 URL: https://issues.apache.org/jira/browse/FLINK-31736
             Project: Flink
          Issue Type: Improvement
          Components: Runtime / REST
            Reporter: David Morávek


This is a follow-up to the [https://github.com/apache/flink/pull/22350#discussion_r1158133570]

 

We only have a per cluster /config endpoint that allows us to enable/hide features on the client side (web UI). Still, sometimes we need to be able to match these features with the Job's capabilities to provide a great experience.

 

An example is that when the rescaling is enabled (and supported by the cluster -> we're using AdaptiveScheduler) and the user submits a batch job, we fallback to a different scheduler (AdaptiveBatchScheduler) that doesn't support rescaling. However, we still show the rescaling buttons in the WebUI.

 

Another example of Job's capabilities could be whether it supports checkpointing/savepoints.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)