[
https://issues.apache.org/jira/browse/FLINK-27060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17522699#comment-17522699
]
ConradJam commented on FLINK-27060:
-----------------------------------
I think this is a good proposal, and if we implement this feature, we can
deploy the JAR without having to repackage it every time we change the job
configuration. If you have implemented this function, could you please attach
the address for my reference [~Zhanghao Chen]
> Extending /jars/:jarid/run API to support setting Flink configs
> ---------------------------------------------------------------
>
> Key: FLINK-27060
> URL: https://issues.apache.org/jira/browse/FLINK-27060
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / REST
> Reporter: Zhanghao Chen
> Priority: Major
>
> *Background*
> Users want to submit job via Flink REST API instead of Flink CLI which is
> more heavy-weight in certain scenarios, for example, a lightweight data
> processing workflow system that has Flink related systems.
> Currently, the /jars/:jarid/run API
> ([https://nightlies.apache.org/flink/flink-docs-master/docs/ops/rest_api/#jars-jarid-run)]
> only supports a few selected Flink config options listed in the doc
> (parallelism、savepoint path and allow non-restored state), which is
> insufficient for practical use.
> *Proposed Changes*
> Extending the /jars/:jarid/run API with an additional request body parameter
> "configs", which is a map of flink configuration option-value pairs set by
> users.
> For backward compatibility, we can retain the existing body parameters like
> "allowNonRestoredState", and when there's conflicting configurations, let the
> values set explicitly with existing body parameters take higher precedence
> over the values set by configs.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
