Add option to round up CPU quota

[waltherlee]

Add a RoundQuotaFunc option to control how the CPU quota should be rounded. The default is still rounding down (floor).
This is related to issue #78

[CLAassistant]

All committers have signed the CLA.

[waltherlee]

I just noticed there's a similar PR. This one's also related to the comments in #68

[abhinav]

Thanks for making this change, @wallee94!
I think the maintainers might find this change more digestible.
I've left a comment about one possible improvement, but I'll defer to the maintainers on it.

CC @r-hang @sywhang @mway

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (c9adbb9) 93.19% compared to head (b686bfa) 93.96%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #79      +/-   ##
==========================================
+ Coverage   93.19%   93.96%   +0.77%     
==========================================
  Files           9       10       +1     
  Lines         338      348      +10     
==========================================
+ Hits          315      327      +12     
+ Misses         19       16       -3     
- Partials        4        5       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[waltherlee]

Added a few tests for the new lines in runtime. I don't have linux so I couldn't run them on my local. I'll have to wait for the checks here 👀

[sandyskies]

@sywhang

[abhinav]

LGTM save minor edits and a suggested fix for being able to use the RoundQuota option.
I'm not a maintainer, though, so I'll defer to @sywhang @r-hang and @mway.

[waltherlee]

I finally set up a linux machine to run those tests! I also changed the implementation to take a round function in the option instead of setting the operation with a const value. I felt it was simpler to use and maintain.

[r-hang]

looks good to me.

Does it make sense to test the round == nil case within internal/runtime since that's newly added?

[waltherlee]

@r-hang Sure! Added a test to check that the value is rounded down then the round arg is nil