[ https://issues.apache.org/jira/browse/ZOOKEEPER-744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12871479#action_12871479 ]
Travis Crawford commented on ZOOKEEPER-744: ------------------------------------------- Looks awesome! One nit -- many monitoring systems do not interpret strings so it may be appropriate to export everything as numbers. For example, consider a script that loops through these values poking them into Ganglia (or other timeseries database). The script would need special-cased to handle "leader". Later, as more values are added the import script would need updated with the new strings. Doing everything as numbers ensures new values would "just work" without updating other systems. With that in mind, perhaps: zk_server_state 1 (instead of: leader) zk_build_timestamp unix_timestamp (instead of build string) > Add monitoring four-letter word > ------------------------------- > > Key: ZOOKEEPER-744 > URL: https://issues.apache.org/jira/browse/ZOOKEEPER-744 > Project: Zookeeper > Issue Type: New Feature > Components: server > Reporter: Travis Crawford > Attachments: zk-ganglia.png, ZOOKEEPER-744.patch > > > Filing a feature request based on a zookeeper-user discussion. > Zookeeper should have a new four-letter word that returns key-value pairs > appropriate for importing to a monitoring system (such as Ganglia which has a > large installed base) > This command should initially export the following: > (a) Count of instances in the ensemble. > (b) Count of up-to-date instances in the ensemble. > But be designed such that in the future additional data can be added. For > example, the output could define the statistic in a comment, then print a key > "space character" value line: > """ > # Total number of instances in the ensemble > zk_ensemble_instances_total 5 > # Number of instances currently participating in the quorum. > zk_ensemble_instances_active 4 > """ > From the mailing list: > """ > Date: Mon, 19 Apr 2010 12:10:44 -0700 > From: Patrick Hunt <ph...@apache.org> > To: zookeeper-u...@hadoop.apache.org > Subject: Re: Recovery issue - how to debug? > On 04/19/2010 11:55 AM, Travis Crawford wrote: > > It would be a lot easier from the operations perspective if the leader > > explicitly published some health stats: > > > > (a) Count of instances in the ensemble. > > (b) Count of up-to-date instances in the ensemble. > > > > This would greatly simplify monitoring& alerting - when an instance > > falls behind one could configure their monitoring system to let > > someone know and take a look at the logs. > That's a great idea. Please enter a JIRA for this - a new 4 letter word > and JMX support. It would also be a great starter project for someone > interested in becoming more familiar with the server code. > Patrick > """ -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.