John O'Hara

The calibration phase is fixed, the `%` impact of total runtime depends on the duration of the test. wrk/wrk2 will not compile to native as-is. They are just predefined benchmarks...

This can be caused if there is no remote service listening on a port, or the service is returning a response that is not valid, e.g. ``` #1 Sequence request(0)(4/4),...

An example where determining steady-state in post-processing is sub-optimal is if you want to run the same benchmark against multiple implementations of the same system under test (possibly using different...

Is happening on 11.0.18 and also if i switch to Shenandoah ``` # # A fatal error has been detected by the Java Runtime Environment: # # SIGSEGV (0xb) at...

Hi @helio-frota, thanks for reporting this issue. I am going to re-open as I think the error message reported to users could be more intuitive than a stack trace

no problem

> Why would we have the N+1 problem? Sry, read comment before coffee, N+1 comes from the fact we need to scan first to see if sub object exists.

I like the idea of **not** providing grafite dialect support directly into the api. To determine if arrays are arrays of value options vs label values, could we not construct...

I wouldn't worry about the openAPI spec, if it changes, we will just bump to 0.14. We need an api that makes the most sense rationally and does not complicate...

Ah, I now see this issue wrt filtering the api result: https://github.com/Karm/collector/pull/34