You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@zookeeper.apache.org by "Bogdan Kanivets (JIRA)" <ji...@apache.org> on 2018/03/02 07:31:00 UTC

[jira] [Created] (ZOOKEEPER-2990) Implement probabilistic tracing

Bogdan Kanivets created ZOOKEEPER-2990:
------------------------------------------

             Summary: Implement probabilistic tracing
                 Key: ZOOKEEPER-2990
                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2990
             Project: ZooKeeper
          Issue Type: Improvement
          Components: server
            Reporter: Bogdan Kanivets
            Assignee: Bogdan Kanivets


It would be nice to have an ability to do probabilistic tracing similar to Cassandra  [nodetool|https://docs.datastax.com/en/dse/5.1/dse-admin/datastax_enterprise/tools/nodetool/toolsSetTraceProbability.html]

This will help debug issues in prod systems.

I'd like to contribute if everyone is ok with the feature.

My suggestion is to add an extra parameter to ZooTrace to handle it. Questions:
 * should it be one global param or per each ZooTrace mask? I'm thinking per mask
 * should it be a new 4lw or part of 'stmk'? Leaning towards new word and refactoring param passing to words (stmk is a special case right now).
 * there are places in the code that use LOG.trace directly. That will have to change to ZooTrace

I can make some initial implementation for demo/review.

 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)