Call for Help: Flame Graphs and Continuous Integration

[Craig Rodrigues]

Hi,

At MeetBSD, Brendand Gregg at Netflix gave a nice presentation
on Performance Analysis, where he mentioned Flame Graphs:

http://www.slideshare.net/brendangregg/meetbsd2014-performance-analysis
http://www.brendangregg.com/flamegraphs.html

I would like some scripts that do the following:

(1)  Boot a FreeBSD image in a bhyve VM
(2)  Runs the necessary commands to generate the info needed to
      generate a flame graph
(3)  Shutdown the bhyve VM
(4)  Directly mount the bhyve VM after it has been shut down
(5)  Pull out any data files from the VM
(6)  Do any offline analysis of the data files, and generate the flame graph
(7)  Make flame graph accessible online

(In the https://github.com/freebsd/freebsd-ci repository, I already
  have Python scripts to do (1), (3), (4) which is how I run
  kyua, so starting from there would be good).

I would also like:

(A)  Code should be committed to https://github.com/freebsd/freebsd-ci
(B)  Code should run under FreeBSD
(C)  Code should be easy to run under a Jenkins job at
       https://jenkins.freebsd.org
(D)  Feel free to use modern programming languages and libraries.
       I am biased to Python, but am open to anything that works under
       FreeBSD.
(E)  Followup status and discussion should happen on
       the freebsd-testing at freebsd.org mailing list

This would be a good first step to have more freely available
scripts for doing performance analysis of FreeBSD.

I am a bit overstretched at the momemt.  Is there anyone
out there who would be interested in trying to tackle this
stuff and get something implemented?

Thanks.

--
Craig