Sean Hunter created NIFI-16015:
----------------------------------
Summary: Status averageLineageDuration is improperly calculated
for failure queues looped back to SELF
Key: NIFI-16015
URL: https://issues.apache.org/jira/browse/NIFI-16015
Project: Apache NiFi
Issue Type: Bug
Components: Core Framework
Affects Versions: 2.9.0
Reporter: Sean Hunter
Attachments: avgLineageDuration.bug.writeup.md
TL;DR - averageLineageDuration (as reported via status monitoring) is
calculated by taking aggregateLineageMillis and dividing by (flowFilesOut +
flowFilesRemoved). The problem is that flowFilesOut and flowFilesRemoved
notably exclude flowfiles that are looped back to the same processor, but that
is NOT the case for how aggregateLineageMillis is calculated.
LLM explanation attached.
I have not (yet) created a suggestion for how to resolve as I don't believe I
have enough context to know if this data is being used elsewhere in NiFi (so
could break things) or by other users (so could break based on their
expectations).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
