Adam/Eric, thanks for your helpful responses.
Eric, I'm looking at your openTSDB port. With it and the tcollector changes, if I understand correctly, I can use an existing Accumulo backend to store the metrics ? I wasn't sure because src/create_table.sh still refers to HBase. Also, if there are any special setup instructions instead or in addition to http://opentsdb.net/getting-started.html, would be good to know. -Ameet Kini From: Eric Newton [mailto:[email protected]] Sent: Friday, April 13, 2012 5:31 PM To: [email protected] Subject: Re: monitoring Oops... premature send. Also check out: http://github.com/ericnewton/tcollector -Eric On Fri, Apr 13, 2012 at 5:29 PM, Eric Newton <[email protected]<mailto:[email protected]>> wrote: We have used hobbit, and ganglia... which explains why the master monitoring page shows so much information. I am going to continue to work on opentsdb support. Maybe this weekend I'll get around to teasing apart the hbase/accumulo back-ends and see if they will accept the patches. Check out: http://github.com/ericnewton/opentsdb On Fri, Apr 13, 2012 at 4:05 PM, Kini, Ameet M. <[email protected]<mailto:[email protected]>> wrote: What are folks using to monitor Accumulo - the master status page, Ganglia, openTSDB, anything else? I was looking into Ganglia but wasn't sure how/if Accumulo works with it. Would the Accumulo daemons send metrics directly to the gmond, would the metrics be dumped to a file and then read back, some other indirect way? Any online pointers would also be useful. Thanks, -Ameet Kini
