You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@zeppelin.apache.org by "Theo Sardin (Jira)" <ji...@apache.org> on 2021/05/28 06:50:00 UTC

[jira] [Created] (ZEPPELIN-5394) Kubernetes - Option to have zeppelin kill existing idle interpreter to spawn new one

Theo Sardin created ZEPPELIN-5394:
-------------------------------------

             Summary: Kubernetes - Option to have zeppelin kill existing idle interpreter to spawn new one
                 Key: ZEPPELIN-5394
                 URL: https://issues.apache.org/jira/browse/ZEPPELIN-5394
             Project: Zeppelin
          Issue Type: Wish
          Components: Kubernetes, spark
    Affects Versions: 0.9.0
            Reporter: Theo Sardin


Hello,

 

We have a Zeppelin running on kubernetes, but when too many users are using it / too many notes are ran we end up with too many interpreter pods, eating all the quotas for the kube namespace, and having Kubernetes timeout for all subsequent runs.

 

Would there be any way to have a config options to kill an interpreter that's not running anything at this specific moment in order to spawn any new one ?

 

Thanks and regards 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)