[
https://issues.apache.org/jira/browse/FLINK-27893?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17564401#comment-17564401
]
Gyula Fora commented on FLINK-27893:
------------------------------------
Thank you for the input [~maddisondavid] , I see the value but I am going to
play the devils advocate here a little :)
What you expect to see when you list FlinkDeployments is not necessarily the
state of the job, but the state of the resource. Arguable the job state is part
of it so I am not against showing it but I suggest we show something like:
{noformat}
enum ResourceLifecycleState {
SUSPENDED, UPGRADING, DEPLOYED, STABLE, ROLLING_BACK, ROLLED_BACK, FAILED
}{noformat}
I am working on something like this for tracking detailed metrics about the
operator (measure time in applicaiton lifecycle state transictions).
Not sure UpgradeMode/Savepoint should be shown here, I think thats not really
important as long as it works correctly. Flink version makes sense
> Add additionalPrinterColumns for Job Status and Reconciliation Status in CRDs
> -----------------------------------------------------------------------------
>
> Key: FLINK-27893
> URL: https://issues.apache.org/jira/browse/FLINK-27893
> Project: Flink
> Issue Type: Improvement
> Components: Kubernetes Operator
> Reporter: Jeesmon Jacob
> Assignee: Jeesmon Jacob
> Priority: Major
>
> Right now only columns available when you run "kubectl get flinkdeployment"
> is NAME and AGE. It is beneficial to see the .status.jobStatus.State and
> .status.ReconciliationStatus.State as additionalPrintColumns in CRD. It will
> be beneficial for automation to use those columns for status reporting.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
