> On Nov. 30, 2016, 12:04 a.m., Santhosh Kumar Shanmugham wrote: > > src/test/java/org/apache/aurora/scheduler/sla/SlaUtilTest.java, lines 16-24 > > <https://reviews.apache.org/r/54106/diff/4/?file=1572568#file1572568line16> > > > > Remove extra tabs.
done. > On Nov. 30, 2016, 12:04 a.m., Santhosh Kumar Shanmugham wrote: > > src/test/java/org/apache/aurora/scheduler/sla/SlaUtilTest.java, line 26 > > <https://reviews.apache.org/r/54106/diff/4/?file=1572568#file1572568line26> > > > > Please add 0th and 100th percentile to the testcases to enfore the > > boundary condition. done. > On Nov. 30, 2016, 12:04 a.m., Santhosh Kumar Shanmugham wrote: > > src/test/java/org/apache/aurora/scheduler/sla/SlaUtilTest.java, lines > > 120-127 > > <https://reviews.apache.org/r/54106/diff/4/?file=1572568#file1572568line120> > > > > I am getting different numbers, from a spreadsheet, > > > > p75 = 75 > > p99 = 89.4 > > p999 = 89.94 The contract that the SlaUtil::percentile is kind of opposite. As the java-doc used to say: - * Reports the percentile value from the given list ordered in a non-descending order. - * Example: [30, 60, 70, 90], the 75 percentile is 30 (i.e. 75% of elements are greater). So when calculating 75-percentile we want a number which is smaller than 75% of our sample set. - Reza ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/54106/#review157349 ----------------------------------------------------------- On Nov. 30, 2016, 6:28 a.m., Reza Motamedi wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/54106/ > ----------------------------------------------------------- > > (Updated Nov. 30, 2016, 6:28 a.m.) > > > Review request for Aurora, David McLaughlin, Joshua Cohen, and Santhosh Kumar > Shanmugham. > > > Repository: aurora > > > Description > ------- > > __Modification to SlaUtil::percentile to compute percentiles by > interpolation__ > > The calculation of mttX (median-time-to-X) depends on the computation of > percentile values. The current implementation does not behave nicely with a > small sample size. For instance, for a given sample set of {50, 150}, > 50-percentile is reported to be 50. Although, 100 seems a more appropriate > return value. > > This patch uses Guava's Quantile for interpolation. > > > Diffs > ----- > > src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java > 263647e802b664fd3ea21cef8109fb7a9252356f > src/main/java/org/apache/aurora/scheduler/sla/SlaUtil.java > 156b9c0a2fa0c0ec4b7220d5ec2cc40c3e59d1d6 > src/test/java/org/apache/aurora/scheduler/sla/SlaAlgorithmTest.java > eca1beebc906d3354d8c294ba466bc904bde98fa > src/test/java/org/apache/aurora/scheduler/sla/SlaUtilTest.java PRE-CREATION > > Diff: https://reviews.apache.org/r/54106/diff/ > > > Testing > ------- > > Tests added and modified to match the continuous distribution. > I also exported the tests into excel and check the reports percentiles. > > > Thanks, > > Reza Motamedi > >
