[ 
https://issues.apache.org/jira/browse/NIFI-15862?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18116594#comment-18116594
 ] 

ASF subversion and git services commented on NIFI-15862:
--------------------------------------------------------

Commit d352f840c513ae8cff564f6e65a1d8bf0eb5bf05 in nifi's branch 
refs/heads/main from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=d352f840c51 ]

NIFI-15862 Added Virtual Thread Processor Scheduling Strategy (#11164)

Co-authored-by: Cursor <[email protected]>
Signed-off-by: David Handermann <[email protected]>

> Move Processors to Virtual Threads
> ----------------------------------
>
>                 Key: NIFI-15862
>                 URL: https://issues.apache.org/jira/browse/NIFI-15862
>             Project: Apache NiFi
>          Issue Type: Improvement
>          Components: Core Framework
>            Reporter: Mark Payne
>            Assignee: Mark Payne
>            Priority: Major
>          Time Spent: 7h 40m
>  Remaining Estimate: 0h
>
> When we moved to Java 21 for NiFi 2.0 one of the exciting capabilities that 
> we discussed was the ability to move to virtual threads for scheduling 
> processors. This provides a handful of benefits:
>  * The scheduling code becomes much simpler
>  * It's a bit more efficient, with no need to constantly schedule tasks and 
> then cancel futures in the case of a yield, etc.
>  * We could use ThreadLocal whereas right now it does not work well
>  * Use of Logging MDC could be improved
>  * Thread dumps are much clearer because we pin a given 'task' to a single 
> thread. So things like AbstractProcessor calling ProcessSession.commit can be 
> tracked back to a specific thread and therefore a specific processor, etc.



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

Reply via email to