You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Kevin Burton <bu...@spinn3r.com> on 2015/05/30 01:06:39 UTC
Tracing ActiveMQ requests...
I’m trying to track down performance issues for our broker.
I wrote a quick stopwatch around sending messages and it’s taking 2-6
seconds to send requests to ActiveMQ.
I have NO idea why this could be because I have reasonable CPU on this box
and what could be happening. Our broker runs out of memory, no disk is
involved, plenty of CPU and network, etc.
It would be nice if we had more tracing support in ActiveMQ.
If you’ve ever used Cassandra’s tracing support you’ll know what I’m
talking about
ALA dapper-style tracing.
http://research.google.com/pubs/pub36356.html
Essentially it would track each step that could potentially take time and
print them out as they happen.
Maybe one way this could be accomplished is to send messages to a dedicated
queue for each trace attempt?
Either that or log the trace in the log file.
This would NOT be log4j tracing.. which are just log lines. This would be
a list of events and timings.
Maybe this would be better to add to ActiveMQ 6.0 though.. but I guess we
can figure it out as time goes by.
Thoughts?
--
Founder/CEO Spinn3r.com
Location: *San Francisco, CA*
blog: http://burtonator.wordpress.com
… or check out my Google+ profile
<https://plus.google.com/102718274791889610666/posts>
Re: Tracing ActiveMQ requests...
Posted by Kevin Burton <bu...@spinn3r.com>.
Oh.. nice. well I didn’t see that.
I ended up implementing something with advisories to trace message flow.
It’s actually interesting as that you can just connect to a broker and
trace message flow.
We might want to open source it at some point.
On Fri, Jun 5, 2015 at 6:40 AM, Gary Tully <ga...@gmail.com> wrote:
> have you run with trace=true on the transport urls. this will log4j
> trace each openwire command on a per connection logger so that you can
> at least identify the actual command that is taking time.
> connection creation. producer creation, message send etc.
>
> On 30 May 2015 at 00:06, Kevin Burton <bu...@spinn3r.com> wrote:
> > I’m trying to track down performance issues for our broker.
> >
> > I wrote a quick stopwatch around sending messages and it’s taking 2-6
> > seconds to send requests to ActiveMQ.
> >
> > I have NO idea why this could be because I have reasonable CPU on this
> box
> > and what could be happening. Our broker runs out of memory, no disk is
> > involved, plenty of CPU and network, etc.
> >
> > It would be nice if we had more tracing support in ActiveMQ.
> >
> > If you’ve ever used Cassandra’s tracing support you’ll know what I’m
> > talking about
> >
> > ALA dapper-style tracing.
> >
> > http://research.google.com/pubs/pub36356.html
> >
> > Essentially it would track each step that could potentially take time and
> > print them out as they happen.
> >
> > Maybe one way this could be accomplished is to send messages to a
> dedicated
> > queue for each trace attempt?
> >
> > Either that or log the trace in the log file.
> >
> > This would NOT be log4j tracing.. which are just log lines. This would
> be
> > a list of events and timings.
> >
> > Maybe this would be better to add to ActiveMQ 6.0 though.. but I guess we
> > can figure it out as time goes by.
> >
> > Thoughts?
> >
> > --
> >
> > Founder/CEO Spinn3r.com
> > Location: *San Francisco, CA*
> > blog: http://burtonator.wordpress.com
> > … or check out my Google+ profile
> > <https://plus.google.com/102718274791889610666/posts>
>
--
Founder/CEO Spinn3r.com
Location: *San Francisco, CA*
blog: http://burtonator.wordpress.com
… or check out my Google+ profile
<https://plus.google.com/102718274791889610666/posts>
Re: Tracing ActiveMQ requests...
Posted by Gary Tully <ga...@gmail.com>.
have you run with trace=true on the transport urls. this will log4j
trace each openwire command on a per connection logger so that you can
at least identify the actual command that is taking time.
connection creation. producer creation, message send etc.
On 30 May 2015 at 00:06, Kevin Burton <bu...@spinn3r.com> wrote:
> I’m trying to track down performance issues for our broker.
>
> I wrote a quick stopwatch around sending messages and it’s taking 2-6
> seconds to send requests to ActiveMQ.
>
> I have NO idea why this could be because I have reasonable CPU on this box
> and what could be happening. Our broker runs out of memory, no disk is
> involved, plenty of CPU and network, etc.
>
> It would be nice if we had more tracing support in ActiveMQ.
>
> If you’ve ever used Cassandra’s tracing support you’ll know what I’m
> talking about
>
> ALA dapper-style tracing.
>
> http://research.google.com/pubs/pub36356.html
>
> Essentially it would track each step that could potentially take time and
> print them out as they happen.
>
> Maybe one way this could be accomplished is to send messages to a dedicated
> queue for each trace attempt?
>
> Either that or log the trace in the log file.
>
> This would NOT be log4j tracing.. which are just log lines. This would be
> a list of events and timings.
>
> Maybe this would be better to add to ActiveMQ 6.0 though.. but I guess we
> can figure it out as time goes by.
>
> Thoughts?
>
> --
>
> Founder/CEO Spinn3r.com
> Location: *San Francisco, CA*
> blog: http://burtonator.wordpress.com
> … or check out my Google+ profile
> <https://plus.google.com/102718274791889610666/posts>