Fix StatsD p100 calculation

[tpounds]

Fixes an index out of range panic when calculating p100. The following line evaluates to the length of the array when n = 100.

i := int(float64(len(rs.perc)) * float64(n) / float64(100))

I've updated the existing tests to check p0 and p100 boundary conditions.

Required for all PRs:

• Signed CLA.
• Associated README.md updated. (N/A)
• Has appropriate unit tests.

[tpounds]

@danielnelson Looks like CircleCI failures are unrelated. LMK if there is something I should address.

[danielnelson]

Yeah, AWS is having some issues I think, I'll try to re-run the tests later today.

[danielnelson]

I changed the destination branch to master, you will need to rebase with -i and remove the unrelated changes and force push.

[tpounds]

Rebased. CircleCI failure still appears to be unrelated.