Benchmark tests for process*(...)

[JensRantil]

I wrote a couple of benchmarks for statsdaemon and thought that they maybe could be of interest to enable people to test throughput and potential speed improvements. Let me know whether you find them useful or not.

[mreiferson]

these look good, thanks!

mind posting the output of the benchmark on this PR for posterity?

[JensRantil]

You're welcome!

mind posting the output of the benchmark on this PR for posterity?

Sure!

$ go test -bench=".*"
PASS
BenchmarkManyDifferentSensors          1    2014624926 ns/op
BenchmarkOneBigTimer           1    1958929096 ns/op
BenchmarkLotsOfTimers          1    2012189938 ns/op
ok      github.com/bitly/statsdaemon    12.082s

I've also played around with making the processing more concurrent. It's yielded a speedup, but at the cost of code readability. If you are curious, you can have a look here.

[mreiferson]

Thanks. Feel free to open up a separate PR for the performance improvements so we can take a look.

This PR just needs a rebase from the recent merges...

[JensRantil]

Thanks. Feel free to open up a separate PR for the performance improvements so we can take a look.

Will do in a sec.

This PR just needs a rebase from the recent merges...

Rebase done. Should be mergable now.

[JensRantil]

Will do in a sec.

I seem to have a couple of conflicts. Will submit hopefully tomorrow.

[mreiferson]

looks like travis is reporting test failures on this branch...

[JensRantil]

looks like travis is reporting test failures on this branch...

Fixed. My bad. Usually I get an e-mail when Travis is failing. Not sure why I didn't this time.