[rabbitmq-discuss] Web Management Plugin Sluggishness

Mon Oct 22 11:06:43 BST 2012

On 19/10/12 19:25, Dave Seltzer wrote:
> Seems like the big values are:
> {490415,{rabbit at queue01,os_pid}}
> {1152028,{rabbit at queue01,disk_free}}
> {1694075,{rabbit at queue02,disk_free}}
> {635870,{rabbit at queue01,disk_free_alarm}}
> {310155,{rabbit at queue02,disk_free_alarm}}
> {1286885,{rabbit at queue01,proc_total}}
> {960135,{rabbit at queue02,statistics_level}}
> {956790,{rabbit at queue02,erlang_version}}
> {949552,{rabbit at queue01,run_queue}}
> {93713,{rabbit at queue02,run_queue}}
> {974491,{rabbit at queue01,processors}}

While some of those involve doing real work (e.g. disk_free), many of 
them (os_pid, erlang_version, processors) are really trivial lookups. So 
I suspect it's the cost of the RPC call that is dominating.

I assume that the server is fairly busy; if so then the RPC calls may 
just be taking some time to get scheduled. Although if they're taking 
over a second that's quite busy!

But this suggests that there's probably not an easy fix I'm afraid - 
apart from wait for 3.0 where this is solved.

Cheers, Simon

-- 
Simon MacMullen
RabbitMQ, VMware