[
https://issues.apache.org/jira/browse/NIFI-611?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Mark Payne updated NIFI-611:
----------------------------
Description:
Would like to know metrics about each of the repositories: Content, FlowFile,
Provenance.
- Number of session commits
- Number of Prov Events added to repository, how long it took (CPU time, not
wall time)
- Number of Prov Events Index; how long it took (CPU time, not wall time)
- Number of FlowFiles "processed" (created/added to flow, removed from flow,
expired, transferred to another connection)
- Number of times Prov Repo caused backpressure; how long did it last (Wall
time, not CPU time)
- Average number of records per update
Per Component metrics:
- Number of updates/second occurring
- Average duration of repository update (CPU time)
- Average number of records per update
These should be included per component, and that can then be extrapolated out
to the group level.
was:
Would like to know metrics about each of the repositories: Content, FlowFile,
Provenance.
How many updates/second are occurring per processor/component? Average time?
Average number of records per update?
Would also like to know:
* how many FlowFiles per Content Claim on avg.
* how much time are we spending checking for work when there is none?
This should be included per component, and that can then be extrapolated out to
the group level.
> Add additional metrics to those reported by NiFi
> ------------------------------------------------
>
> Key: NIFI-611
> URL: https://issues.apache.org/jira/browse/NIFI-611
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Reporter: Mark Payne
> Fix For: 0.4.0
>
>
> Would like to know metrics about each of the repositories: Content, FlowFile,
> Provenance.
> - Number of session commits
> - Number of Prov Events added to repository, how long it took (CPU time, not
> wall time)
> - Number of Prov Events Index; how long it took (CPU time, not wall time)
> - Number of FlowFiles "processed" (created/added to flow, removed from flow,
> expired, transferred to another connection)
> - Number of times Prov Repo caused backpressure; how long did it last (Wall
> time, not CPU time)
> - Average number of records per update
> Per Component metrics:
> - Number of updates/second occurring
> - Average duration of repository update (CPU time)
> - Average number of records per update
> These should be included per component, and that can then be extrapolated out
> to the group level.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
