[Oslc-pm] OSLC-PM and Web Applications

John Arwe johnarwe at us.ibm.com
Tue Jan 8 12:01:45 EST 2013


I see nothing in the archives to suggest this was ever responded to.  Was 
it handled on a call?

The Avg/Min/Max/StdDev pattern being repeated makes me wonder out loud if 
this is in reality some form of qualification of the value.  Perf Mon to 
date I don't think considered that to be in scope.  It is easy enough to 
do via extensions, or you can convince (try to) the WG to either 
add/modify in-scope scenarios to require these.

"hits" semantics are unclear to me, so cannot comment on whether they are 
new/map to existing.

Any questions around CRTV needs to be answered by the owners of that 
vocabulary, the Reconciliation working group [1].  Perf Mon uses that only 
as an example of how one could "attach" metrics to types already defined 
elsewhere.


[1] http://open-services.net/wiki/reconciliation/

Best Regards, John

Voice US 845-435-9470  BluePages
Tivoli OSLC Lead - Show me the Scenario


"Oslc-Pm" <oslc-pm-bounces at open-services.net> wrote on 12/11/2012 01:26:55 
PM:

> From: Boris Kuschel <kuschel at ca.ibm.com>
> To: oslc-pm at open-services.net
> Date: 12/11/2012 01:27 PM
> Subject: Re: [Oslc-pm] OSLC-PM and Web Applicaitions
> Sent by: "Oslc-Pm" <oslc-pm-bounces at open-services.net>
> 
> I am currently using the EMS, CRTV and PERFMON specs but I have 
> identified the following caps especially around the web application 
area. 
> 
> *= new
> 
> RequestMetrics
> -> AvgResponseTime*
> -> MaxResponseTime*
> -> MinResponseTime*
> -> StandardDeviation*
> -> AvgPayloadSize*
> -> MaxPayloadSize*
> -> MinPayloadSize*
> -> AvgRequestFailures
> -> Hits*
> -> CpuUsed
> 
> RequestFailures
> -> AvgResponseTime*
> -> MaxResponseTime*
> -> MinResponseTime*
> -> StandardDeviation*
> -> Hits*
> -> CpuUsed
> 
> There is a need for a light weight service (J2EE, OSGi). I have seen
> the CRTV ServiceInstance predicate but the documentation seems to 
> indicate that this a customer related servicing offering.
> 
> ServiceMetrics*
>  ->AvgResponseTime*
> -> MaxResponseTime*
> -> MinResponseTime*
> -> StandardDeviation*
> -> AvgRequestFailures*
> -> Hits*
> -> CpuUsed
> -> parent (count be request or service, crtv?)*
> 
> SQLStatementMetrics*
> -> AvgResponseTime*
> -> MaxResponseTime*
> -> MinResponseTime*
> -> StandardDeviation*
> -> parent (count be request or service, crtv?)*
> 
> Is it small enough? I have other data that deals more with 
> performance of active instances of services and requests as well but
> including the above would be a great first step.
> 
> Boris Kuschel
> Software Architect
> Smarter Server - IBM Software, Rational
> Phone: +1-514-964-1290
> Email: kuschel at ca.ibm.com
> My Blog: http://blog.boriskuschel.com/
> 
> [image removed] John Arwe ---12/07/2012 01:20:55 PM---This use 
> certainly sounds compatible on the surface. With respect to specific
> metrics, I don't know
> 
> From: John Arwe <johnarwe at us.ibm.com>
> To: oslc-pm at open-services.net, 
> Date: 12/07/2012 01:20 PM
> Subject: Re: [Oslc-pm] OSLC-PM and Web Applicaitions
> Sent by: "Oslc-Pm" <oslc-pm-bounces at open-services.net>
> 
> 
> 
> This use certainly sounds compatible on the surface. 
> With respect to specific metrics, I don't know of any specific 
> scenarios in-scope for this spec version.  Have you done a gap 
> analysis, e.g. tried to lay out that data using the current draft to
> see what does not fit obviously?  If the gap is considered "small 
> enough", the WG might conceivably elect to add scenario(s) in this 
> version and cover them.  If it's not, there's always the next spec 
> version as a catcher for the scenarios, assuming they get 
> prioritized highly enough by the WG. 
> It might be that terms from other vocabularies are needed (you might
> look at CRTV under the Reconciliation WG), but that's standard 
> procedure not an impediment of any kind. 
> Best Regards, John
> 
> Voice US 845-435-9470  BluePages 
> Tivoli OSLC Lead - Show me the Scenario 
> 
> 
> 
> 
> From:        Boris Kuschel <kuschel at ca.ibm.com> 
> To:        oslc-pm at open-services.net 
> Date:        11/30/2012 04:18 PM 
> Subject:        [Oslc-pm] OSLC-PM and Web Applicaitions 
> Sent by:        "Oslc-Pm" <oslc-pm-bounces at open-services.net> 
> 
> 
> 
> HI,
> 
> We are intending to using the OSLC-PM (and associated specs) to 
> provide performance data for a OSGi-based web application (the same 
> concepts apply to J2EE). I notice that there is a lot of metrics 
> relating to OS, a few about JVM (more around the specifics of the GC
> would be nice, nursery, perm space,etc.) but there is general lack 
> of metric details regarding web requests. There are a few tempting 
> predicates (Process, ServiceInstance) that almost hit the mark but not 
quite.
> 
> Consider the following scenario: I have a system running a web 
> application for which I wish to understand health status. There is a
> general slow down in the system and I expect to be able to 
> understand where the slowdown is occurring. I would like to be able 
> to pull down metrics information and see the following:
> 
> Active/Most consuming requests (or some amount):
> ->User Session, user, age, language, agent
> -->HTTP Request: type, time spent, cpu cycles consumed, bytes sent, 
> bytes received
> ---> Service Invoked: time spent, cpu cycles consumed
> -----> Activities (SQL, triple store, external service, etc.): 
> description, time spent, cpu cycles consumed
> ------> Services Invoked... etc.
> 
> Is there consideration to make this information as part of the OSLC-
> PM vocabulary?
> 
> Thanks,
> 
> Boris Kuschel
> Software Architect
> Smarter Server - IBM Software, Rational
> _______________________________________________
> Oslc-Pm mailing list
> Oslc-Pm at open-services.net
> http://open-services.net/mailman/listinfo/oslc-pm_open-services.net
> _______________________________________________
> Oslc-Pm mailing list
> Oslc-Pm at open-services.net
> http://open-services.net/mailman/listinfo/oslc-pm_open-services.net
> _______________________________________________
> Oslc-Pm mailing list
> Oslc-Pm at open-services.net
> http://open-services.net/mailman/listinfo/oslc-pm_open-services.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://open-services.net/pipermail/oslc-pm_open-services.net/attachments/20130108/0670d1e8/attachment-0003.html>


More information about the Oslc-Pm mailing list