Hi Lance Ju,

Which collector you are trying to integrate with jMeter ?
As we can see in saga code "Kamon" metrics with StatsD collector is there.
Point[2] in your mail: can we add kamon & StatsD/grafana along with custom 
jMeter image ?

Ref: 
https://github.com/apache/incubator-servicecomb-saga/tree/master/saga-performance


Thanks & Regards
-Mahesh Raju S
Github: maheshrajus

-----Original Message-----
From: Lance Ju [mailto:[email protected]] 
Sent: 10 August 2018 09:32
To: [email protected]
Cc: Sukesh A C <[email protected]>; Mohammad Asif Siddiqui 
<[email protected]>; Vinay Murudi <[email protected]>; Zhenhua 
Hu <[email protected]>
Subject: Re: [DISCUSS] Performance benchmarking tool finding for Saga

Hi, team,
  I've made some progress on this, we can now deploy demos(currently only 
spring-demo supported) into a k8s cluster, then start a jmeter container for 
load tests.
  Some further works are under process:
  1. A jmeter test result collector in the service
  2. A custom jmeter image that upload the result to the collector

  After that we can always check the collected results in the collector, and 
link the load tests to CI/CD systems, to make the test itself 'Continuous'.


Willem Jiang <[email protected]> 于2018年8月9日周四 下午11:40写道:

> FYI.
>
> There are some performance test work[1] we did laster year, that you 
> may take a look.
>
>
> https://github.com/apache/incubator-servicecomb-saga/tree/master/saga-
> performance
>
>
> Willem Jiang
>
> Twitter: willemjiang
> Weibo: 姜宁willem
>
> On Thu, Aug 9, 2018 at 10:48 AM, Willem Jiang <[email protected]>
> wrote:
>
> > Hi
> >
> > The two articles are pretty much same, so you just need to take a 
> > look at one. As they didn't open source the performance test code, 
> > we need to
> setup
> > the performance test based on the demo ourself.
> >
> > Now Ju Zheng is working on build test deploying environment for the 
> > demo application and saga into k8s on this task[1].
> > So the next important work is we should know about the slowest parts 
> > of our system.
> >
> > We did some performance turning for the old saga implementation.  
> > Here
> are
> > some informations that you can take a look.
> > 1. We use the kamon[2] to instrument the saga application, so we can 
> > get the metrics and tracing by applying the annotations of kamon.
> > You can find the example code by searching @EnableKamon in the 
> > saga-core module.
> >
> > 2.  We need to send the metrics to reporter to help us to observer 
> > the whole system.
> > kamon supports lots of reporters, using the InfluxDB and Grafana 
> > could be a good way to go.
> >
> >
> > Please let me know if you have any question about it.
> >
> > [1]https://issues.apache.org/jira/browse/SCB-823
> > [2]http://kamon.io/documentation/get-started/
> >
> >
> > Willem Jiang
> >
> > Twitter: willemjiang
> > Weibo: 姜宁willem
> >
> > On Wed, Aug 8, 2018 at 11:37 PM, mahesh somalaraju < 
> > [email protected]> wrote:
> >
> >> Dear Servicecomb Dev team,
> >>
> >> I am looking forward to contribute the project 
> >> "incubator-servicecomb-saga".
> >> Currently I am understanding the use cases and code flows of Saga.
> >> I have gone through the some sample demos and it is interesting.
> >>
> >> I am looking for finding one performance tool which can calculate 
> >> the performance of saga.
> >> There is no direct tool which can calculate the performance of 
> >> distributed transaction framework.
> >>
> >> I have gone through some links presented in web. These links in 
> >> Chinese language and I translated and gone through.
> >> https://cloud.tencent.com/developer/news/280262
> >> https://blog.csdn.net/aioria_csdn/article/details/81139958
> >>
> >> I am planning to calculate the performance on below performance factors.
> >>
> >> 1)      TPS
> >>
> >> 2)      Response time and TPS when multiple clients presents.
> >>
> >> 3)      Response time and average TPS in multiple thread scenario.
> >>
> >> Any inputs are welcome !  Thank you in advance.
> >>
> >> Thanks & Regards
> >> Mahesh Raju Somalaraju
> >> Gtihub: maheshrajus
> >>
> >>
> >>
> >
>

Reply via email to