You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Yang Wang (Jira)" <ji...@apache.org> on 2022/04/01 05:24:00 UTC

[jira] [Created] (FLINK-26972) Support spot instance in Flink Kubernetes operator

Yang Wang created FLINK-26972:
---------------------------------

             Summary: Support spot instance in Flink Kubernetes operator
                 Key: FLINK-26972
                 URL: https://issues.apache.org/jira/browse/FLINK-26972
             Project: Flink
          Issue Type: New Feature
          Components: Kubernetes Operator
            Reporter: Yang Wang


From the discussion in the ML[1][2], some users are running Flink on spot EC2. Since cloud Kubernetes also supports spot instance[3], it will be great if the K8s operator could be aware of eviction events. Then it could trigger a savepoint and delete the pods gracefully.

 

Maybe this is unnecessary when the Kubernetes HA enabled. Let's keep this ticket open and see what could we do in the Flink K8s operator to make stateful Flink application running on spot instance better.

 

[1]. [https://lists.apache.org/thread/k13mh1c01zsd59kchw4dv160qx4s6l66]

[2]. [https://lists.apache.org/thread/c1x1ys3cdvhd9r4v08z399kch3o34nf0]

[3]. [https://docs.aws.amazon.com/eks/latest/userguide/managed-node-groups.html]



--
This message was sent by Atlassian Jira
(v8.20.1#820001)