You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by "Jay Kreps (Commented) (JIRA)" <ji...@apache.org> on 2011/10/22 23:04:32 UTC
[jira] [Commented] (KAFKA-166) Create a tool to jump JMX data to a
csv file to help build out performance tests
[ https://issues.apache.org/jira/browse/KAFKA-166?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13133480#comment-13133480 ]
Jay Kreps commented on KAFKA-166:
---------------------------------
Added draft patch.
jkreps-mn:kafka-trunk jkreps$ bin/kafka-run-class.sh kafka.tools.JmxTool --help
Option Description
------ -----------
--help Print usage information.
--jmx-url <service-url> The url to connect to to poll JMX
data. See http://http://download.
oracle.
com/javase/7/docs/api/javax/management/remote/JMXServiceURL.
html for details. (default: service:
jmx:rmi:///jndi/rmi://:9999/jmxrmi)
--object-name <name> A JMX object name to query. If no
objects are specified all objects
will be queried.
--reporting-interval <Integer: ms> Interval in MS with which to poll jmx
stats. (default: 5000)
Example:
jkreps-mn:kafka-trunk jkreps$ bin/kafka-run-class.sh kafka.tools.JmxTool --object-name 'kafka:*' --reporting-interval 1000
"time", "kafka:type=kafka.KafkaLog4j:priority", "kafka:type=kafka.SocketServerStats:NumFetchRequests", "kafka:type=kafka.SocketServerStats:NumProduceRequests", "kafka:type=kafka.SocketServerStats:FetchRequestsPerSecond", "kafka:type=kafka.SocketServerStats:TotalBytesRead", "kafka:type=kafka.KafkaLog4j:appender=stdout", "kafka:type=kafka.SocketServerStats:BytesWrittenPerSecond", "kafka:type=kafka.SocketServerStats:AvgProduceRequestMs", "kafka:type=kafka.KafkaLog4j:name", "kafka:type=kafka.SocketServerStats:MaxFetchRequestMs", "kafka:type=kafka.SocketServerStats:BytesReadPerSecond", "kafka:type=kafka.SocketServerStats:TotalProduceRequestMs", "kafka:type=kafka.SocketServerStats:ProduceRequestsPerSecond", "kafka:type=kafka.SocketServerStats:MaxProduceRequestMs", "kafka:type=kafka.SocketServerStats:TotalFetchRequestMs", "kafka:type=kafka.SocketServerStats:TotalBytesWritten", "kafka:type=kafka.SocketServerStats:AvgFetchRequestMs"
1319317539516, INFO, 0, 0, -0.0, 0, log4j:appender=stdout, 0.0, 0.0, root, 0.0, 0.0, 0, -0.0, 0.0, 0, 0, 0.0
1319317540523, INFO, 0, 0, -0.0, 0, log4j:appender=stdout, 0.0, 0.0, root, 0.0, 0.0, 0, -0.0, 0.0, 0, 0, 0.0
1319317541523, INFO, 0, 0, -0.0, 0, log4j:appender=stdout, 0.0, 0.0, root, 0.0, 0.0, 0, -0.0, 0.0, 0, 0, 0.0
1319317542522, INFO, 0, 0, -0.0, 0, log4j:appender=stdout, 0.0, 0.0, root, 0.0, 0.0, 0, -0.0, 0.0, 0, 0, 0.0
1319317543521, INFO, 0, 0, -0.0, 0, log4j:appender=stdout, 0.0, 0.0, root, 0.0, 0.0, 0, -0.0, 0.0, 0, 0, 0.0
...
> Create a tool to jump JMX data to a csv file to help build out performance tests
> --------------------------------------------------------------------------------
>
> Key: KAFKA-166
> URL: https://issues.apache.org/jira/browse/KAFKA-166
> Project: Kafka
> Issue Type: New Feature
> Components: core
> Affects Versions: 0.8
> Reporter: Jay Kreps
> Assignee: Jay Kreps
> Attachments: KAFKA-166.patch
>
>
> In order to get sane performance stats we need to be able to integrate the values we keep in JMX. To enable this it would be nice to have a generic tool that dumped JMX stats to a csv file. We could use this against the producer, consumer, and broker to collect kafka metrics while the tests were running.
--
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