I think DSL has 2 important advantages: - It is more readable in source repositories. - it is better for developers who load test their application in CI/CD. This is clearly an important move JMeter must follow.
so based on the feedbacks of the roadmap thread and this one, I think it must be in the roadmap. If possible, we could start thinking in this thread about: - its syntax, knowing we have an example which I find rather nice, easy and powerful with ruby-jmeter - the language it should be developed in. My personal preference would go for Groovy but I am not an expert in DSLs. - how to express what today we can do through JSR223 languages which might be the most complex think to translate That's for my 2 cents Regards On Thursday, July 28, 2016, Andrey Pokhilko <[email protected]> wrote: > This is one of the big questions for me, why JMeter is so popular, while > gatling/locust and all other similar tools have relaively narrow > audience. I don't have exact answer for this question. > > I think it is a lot because of JMeter UI, which makes it very easy to > start load testing for people who are not familiar with coding at all. > So the weak part of JMeter is apparently it's strongest win. > > Andrey Pokhilko > > On 07/28/2016 10:44 PM, [email protected] <javascript:;> wrote: > > > > -- Changing subject to separate DSL discussion from roadmap thread – > > > > > > > > Andrey, I’m interested in hearing why /you/ think Gatling didn’t > > become more popular. > > > > > > > > RaGe > > > > > > > > > > > > *From: *Andrey Pokhilko <mailto:[email protected] <javascript:;>> > > *Sent: *Thursday, July 28, 2016 11:27 AM > > *To: *[email protected] <javascript:;> <mailto:[email protected] > <javascript:;>> > > *Subject: *Re: Roadmap 2016/2017 > > > > > > > > Gatling tool provide the "code" way to create the tests. Why didn't it > > > > become dominant tool in Open Source field while JMeter with UI did? > > > > > > > > Andrey Pokhilko > > > > > > > > On 07/28/2016 06:08 PM, Richard Friedman wrote: > > > > > JMeter DSL > > > > > - A DSL in which folks could build load tests would go a long way to > > enable > > > > > JMeter in the development lifecycle. No one can write tests in the > > current > > > > > JMX file without opening the UI. Dev/QA and CI/CD have changed very > > much > > > > > over the past 5 years. More performance tests could be built and > > integrated > > > > > into the development lifecycle if the barrier to write one goes down. > > > > > > > > > > > > > > > > > > > > > > > > > On Thu, Jul 28, 2016 at 10:06 AM, Antonio Gomes Rodrigues > > <[email protected] <javascript:;>> > > > > > wrote: > > > > > > > > > >> Hi, > > > > >> > > > > >> > > > > >> 2016-07-28 0:25 GMT+02:00 Milamber <[email protected] > <javascript:;>>: > > > > >> > > > > >>> > > > > >>> On 27/07/2016 13:41, Philippe Mouawad wrote: > > > > >>> > > > > >>>> Hello, > > > > >>>> I think we should work on a roadmap for JMeter to ensure: > > > > >>>> - we prioritize some urgent work on it. > > > > >>>> - we give visibility on future of JMeter to users > > > > >>>> > > > > >>>> > > > > >>>> I see the following things from most to less urgent: > > > > >>>> > > > > >>>> - http2 support. We depend on httpclient for this > > > > >>>> > > > > >>> For me, seems very important. +1!!! > > > > >>> > > > > >> +1 > > > > >> > > > > >> > > > > >>> > > > > >>> - possible rework of core architecture to at least introduce a pool > of > > > > >>>> threads or switch to async model allowing us to take advantage of > > async > > > > >> io > > > > >>> Seems a lot of works in the JMeter's design. > > > > >>> > > > > >> I think it's a lot of work too but it will be great for the future > > > > >> > > > > >>> > > > > >>> - start a migration to JavaFX , a good opportunity would be to > replace > > > > >> the > > > > >>>> old browser used for html rendering > > > > >>>> > > > > >>> Hum. A lot of work. > > > > >>> > > > > >> Lots of work and I don't know if JavaFX is the best choice (I am > > not sure > > > > >> that Oracle will invest a lot in it) > > > > >> > > > > >> Why not have an Eclipse RCP client? > > > > >> > > > > >> > > > > >>> > > > > >>> > > > > >>>> As side features: > > > > >>>> - DSL ? > > > > >>>> - JSON format instead of XMLfor jmx plans if dsl is too heavy > > change ? > > > > >>>> - Fix undo /redo feature > > > > >>>> > > > > >>> > > > > >>> Not priority? > > > > >>> > > > > >> I have no opinion on it. > > > > >> > > > > >> Have an interactive report will be great too > > > > >> It will allow to make better diagnostic. > > > > >> > > > > >> For example, use Apache Zeppelin (https://zeppelin.apache.org/) > > will allow > > > > >> this. > > > > >> It can take csv file in input or database (e.g. apache cassandra) and > > > > >> produce great custom graphic > > > > >> > > > > >> Have Thread Groups like JMeter plugins (Stepping Thread Group, > Ultimate > > > > >> Thread Group, Concurrency Thread Group, Arrivals Thread Group, > > Free-Form > > > > >> Arrivals Thread Group) will be great too. > > > > >> Neoload and Loadrunner have it and are used for a lot of testers. > > > > >> > > > > >> > > > > >> Another question, do we want to have an easy installation (like > > today with > > > > >> a zip) or a more complex installation (a database + server + rich > > client)? > > > > >> > > > > >> If you check HP Performance center (Loadrunner) and Gatling > > Frontline, they > > > > >> have : > > > > >> a rich client to script (vugen for HP and an IDE for Gatling) > > > > >> a database to store load test result (cassandra for Gatling) > > > > >> a web interface to run/check/view the load test > > > > >> > > > > >> Same with SaaS JMeter solution > > > > >> > > > > >> This solution is "more enterprise ready" and allow easy > > collaborative work > > > > >> > > > > >> Antonio > > > > >> > > > > >> > > > > >>> > > > > >>> Milamber > > > > >>> > > > > >>> > > > > >>> - ... > > > > >>>> Ideas welcome > > > > >>>> Regards > > > > >>>> Philippe M. > > > > >>>> > > > > >>>> > > > > >>>> > > > > >>>> > > > > > > > > > > > > > > > > > > > > > > -- Cordialement. Philippe Mouawad.
