| Author |
Message
|
| WBIMaestro |
 Posted: Wed Aug 15, 2012 9:32 am Post subject: metering flow usage |
 |
|
Acolyte
Joined: 18 Feb 2005
Posts: 53
|
Hi,
are there any features in the broker or any supporting utilities that would tell me how much CPU memory etc... a particular flow is consuming? basically when setting up a service-oriented architecture, how does one track and prevent/alert on a given flow being overloaded by mesages causing excessive processor and CPU usage? any best practices from the listers here?
thx |
|
| Back to top |
|
 |
| Vitor |
| Posted: Wed Aug 15, 2012 9:38 am Post subject: Re: metering flow usage |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| WBIMaestro wrote: |
| are there any features in the broker or any supporting utilities that would tell me how much CPU memory etc... a particular flow is consuming? |
Yes. Research the statistics and monitoring facilities in the InfoCenter.
| WBIMaestro wrote: |
| basically when setting up a service-oriented architecture, how does one track and prevent/alert on a given flow being overloaded by mesages causing excessive processor and CPU usage? |
You use SOAP over JMS & monitor input queue depth. 
You can also use the facilties I menetion above to track usage in a test environment to scale production correctly, and usage over time to rescale as needed.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| WBIMaestro |
| Posted: Thu Aug 16, 2012 6:29 am Post subject: |
|
|
Acolyte
Joined: 18 Feb 2005
Posts: 53
|
actualy, the question is more along the lines of 'how would i implement governance around a SOA solution based on broker'.
Message flow statistics will give me throughput info, but how about CPU, memory consumption as well as threads being spawned for a specific message flow. I guess I could limit the number of threads for the message flow right? basically i want to be able to detect and alert on scenarios where a runaway requesting app dosent overload a particular message flow or the corr broker and end up denying service or causing an outage to other flows and brokers on that server.. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Thu Aug 16, 2012 6:39 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Aug 16, 2012 6:48 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
You should ideally have a gateway/frontend http server in front of your brokers, to ensure that you can properly isolate your consumer applications from the actual physical addresses of the message flows that are acting as providers (or provider-proxies).
You can monitor the statistics on the broker http listeners and the eg soap listeners. But you should also monitor the gateway traffic.
And in general if you ask about "governance" and "message broker", you'll get told "WSRR". |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Aug 16, 2012 6:51 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| WBIMaestro wrote: |
| Message flow statistics will give me throughput info, but how about CPU, memory consumption as well as threads being spawned for a specific message flow. |
Resource statistics give those.
| WBIMaestro wrote: |
| I guess I could limit the number of threads for the message flow right? |
| WBIMaestro wrote: |
| basically i want to be able to detect and alert on scenarios where a runaway requesting app dosent overload a particular message flow or the corr broker and end up denying service or causing an outage to other flows and brokers on that server.. |
That's all about how you distribute flows and which resources they share.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
