> On Jan 2, 2021, at 12:04 PM, [email protected] <[email protected]> 
> wrote:
> 
> Does anyone have experience running performance tests regularly?  
> We have some benchmarks for Jackson, but they're only run manually right now 
> and I'd like a way to see more detailed performance over time.
> 
> I've got a lightly basement server that could host a VM for running the tests 
> nightly, but I don't really know where to start beyond perhaps setting up 
> Jenkins.
> 
> -Drew

I don’t, but these articles by Mark Price come to mind about his experience at 
LMAX with performance testing:

https://epickrram.blogspot.com/2014/05/performance-testing-at-lmax-part-one.html
https://epickrram.blogspot.com/2014/07/performance-testing-at-lmax-part-two.html
https://epickrram.blogspot.com/2014/08/performance-testing-at-lmax-part-three.html

Could be something applicable in there at least conceptually, though the tools 
may have changed since 2014.

I’d start poking at what level of repeatability you can get from a VM. Before 
you wrestle with Jenkins, it’d be good to know if you can in fact tune a guest 
and host kernel in that setup to have at least one or two cores that won’t get 
pre-empted by the kernel, or throttled, or any of the other pernicious things 
that plague benchmark repeatability. Once you’ve got a pretty repeatable 
benchmark sans latency spikes via taskset or isolcpus or whatever, then perhaps 
getting Jenkins to run a workload with that setup might follow easily enough? 
Mark’s blog has more on that sort of kernel fiddling.

Also, while I’m here, I just saw 
https://www.morling.dev/blog/towards-continuous-performance-regression-testing/ 
pop up in my feed the other day. Could be interesting for future test writing.

-- 
You received this message because you are subscribed to the Google Groups 
"jackson-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jackson-dev/2BF62902-7A40-4AFE-BC29-E3EE5A8F8D0D%40mpierce.org.

Reply via email to