[ 
https://issues.apache.org/jira/browse/KAFKA-385?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Joel Koshy updated KAFKA-385:
-----------------------------

    Attachment: KAFKA-385-v3.patch
                KAFKA-385-v3.patch

Changes over v2:
- Rebased (twice!)
- For the remaining (global) histograms switched to biased histograms.
- Addressed Jay's comments:
  - 1: actually there's a workaround - basically pass through the properties
    to the custom reporter. (I provided an example
    (KafkaCSVMetricsReporter). If JMX operations need to be exposed the
    custom reporter will need to implement an mbean trait that extends from
    KafkaMetricsReporterMBean. I did this to avoid having to implement the
    DynamicMBean interface. Since we now have pluggable reporters I removed
    the KafkaMetrics class and the dependency on metrics-ganglia and
    metrics-graphite.
  - 2: changed the logging statements in KafkaApis to just say producer
    requests/fetch requests.
  - 3: I did a quick test as described above, but couldn't see any
    measurable impact.
  - 4: Added KafkaTimer and a unit test (which I'm thinking of removing as
    it is pretty dumb other than showing how to use it).
  - 5: Got rid of the per-key follower catch up time histogram from
    DelayedProduceMetrics.  Furthemore, meters are inexpensive and the
    per-key caught up follower request meter should be sufficient.

                
> RequestPurgatory enhancements - expire/checkSatisfy issue; add jmx beans
> ------------------------------------------------------------------------
>
>                 Key: KAFKA-385
>                 URL: https://issues.apache.org/jira/browse/KAFKA-385
>             Project: Kafka
>          Issue Type: Bug
>            Reporter: Joel Koshy
>            Assignee: Joel Koshy
>              Labels: bugs
>             Fix For: 0.8
>
>         Attachments: example_dashboard.jpg, graphite_explorer.jpg, 
> KAFKA-385-v1.patch, KAFKA-385-v2.patch, KAFKA-385-v3.patch, KAFKA-385-v3.patch
>
>
> As discussed in KAFKA-353:
> 1 - There is potential for a client-side race condition in the 
> implementations of expire and checkSatisfied. We can just synchronize on the 
> DelayedItem.
> 2 - Would be good to add jmx beans to facilitate monitoring RequestPurgatory 
> stats.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to