[
https://issues.apache.org/jira/browse/NIFI-14797?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18013349#comment-18013349
]
John Joseph commented on NIFI-14797:
------------------------------------
[~mosermw],
Thank you for your suggestion, we tried with 200ms. We were able to achieve 15
OPS
!image-2025-08-12-09-56-40-930.png!
> ControlRate Processor sends fewer FlowFiles than expected in NiFi 2.3
> compared to NiFi 1.28
> -------------------------------------------------------------------------------------------
>
> Key: NIFI-14797
> URL: https://issues.apache.org/jira/browse/NIFI-14797
> Project: Apache NiFi
> Issue Type: Improvement
> Affects Versions: 2.3.0
> Reporter: John Joseph
> Priority: Major
> Attachments: image-2025-07-29-11-00-05-438.png,
> image-2025-07-29-11-00-53-864.png, image-2025-08-12-09-56-40-930.png
>
>
> We recently upgraded from NiFi 1.28 to NiFi 2.3 and noticed that the
> *{{ControlRate}}* processor does not honor the configured rate accurately in
> the new version.
> In our use case, the *{{ControlRate}}* is configured with:
> * Rate: 15 flowfiles/second
> * Rate Controlled Criteria: {{FlowFile Count}}
> * Time Duration: {{1 sec}}
> However, the {*}actual average rate is observed to be lower{*}:
> * 15 configured → only ~12.5 observed
> * 18 configured → only ~15 observed
> In NiFi 1.28, the exact configured rate was achieved. The flow design remains
> unchanged, with 3-4 intermediate processors (all with 0 sec run scheduling
> interval) between {{ControlRate}} and {{{}InvokeHTTP{}}}.
> This appears to be a *behavioral change* in +_2.3_+
> We request the team to:
> * Investigate potential regressions in how {{ControlRate}} behaves in the
> newer version.
> * Provide guidance or improvements to ensure accurate rate control,
> especially in batch/integration use cases.
> +*2.3*+
> !image-2025-07-29-11-00-53-864.png!
>
> +*1.28*+
> !image-2025-07-29-11-00-05-438.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
