I went through the JMX mbeans using JConsole and it seems Accumulo supports
a good set of metrics through JMX.
I'm trying to figure out how to leverage openTSBD for Accumulo. From what
I've read so far, I haven't found an easy way to integrate with Accumulo so
that user can use its functionality very easily.
On Fri, Apr 5, 2013 at 1:34 PM, Eric Newton <[EMAIL PROTECTED]> wrote:
> We actually have JMX support. You need to turn it on, though. We should
> just turn it on by default.
> Because we don't use the JMX api very much, it may need some improvement.
> On Fri, Apr 5, 2013 at 1:29 PM, Supun Kamburugamuva <[EMAIL PROTECTED]
> > Thank you all for the input.
> > I think I understand the requirement for the displaying the graphs.
> > For the openTSBD part I'll go through its documentation and Eric's work
> > see what I can find.
> > If we are going for third party tools for monitoring, may be it is better
> > to provide APIs (like JMX) so that users can plug in any tool.
> > Thanks,
> > Supun..
> > On Fri, Apr 5, 2013 at 12:46 PM, Eric Newton <[EMAIL PROTECTED]>
> > wrote:
> > > > ACCUMULO-594 <https://issues.apache.org/jira/browse/ACCUMULO-594>
> > talks
> > > > about OpenTSBD. Can you please give me some references on why we need
> > > > OpenTSBD?
> > > >
> > >
> > > The monitor provides a quick summary of accumulo stats, a bit of
> > > and the most recent ERROR/WARN log messages.
> > > But it's all just kept in memory and only for an hour or so. It isn't
> > > real monitoring tool at all.
> > > It is nice to have *something* we can depend upon as soon as you get
> > > accumulo running.
> > > I do a lot of accumulo support, and having the monitor lets relatively
> > > unsophisticated users see problems right away.
> > >
> > > But we don't have a lot of time to keep putting more energy into making
> > the
> > > monitor a better monitoring tool.
> > > It would be better to integrate with other monitoring tools.
> > >
> > > OpenTSDB stores and presents time-series data. It can monitor anything
> > and
> > > use a NoSQL back-end database to persist the monitored data.
> > > It can handle monitoring thousands of nodes, each with hundreds of
> > > datapoints.
> > > Perhaps there are better tools out there.
> > >
> > > We would still need some tool for the management of log/event data,
> > >
> > > -Eric
> > >
> > --
> > Supun Kamburugamuva
> > Member, Apache Software Foundation; http://www.apache.org
> > E-mail: [EMAIL PROTECTED]; Mobile: +1 812 369 6762
> > Blog: http://supunk.blogspot.com
Member, Apache Software Foundation; http://www.apache.org
E-mail: [EMAIL PROTECTED]; Mobile: +1 812 369 6762