You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Jiang Yan Xu <ya...@jxu.me> on 2017/05/15 05:25:45 UTC
Re: Review Request 53840: Metric in the allocator to track latency in
running allocations.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/53840/#review174871
-----------------------------------------------------------
src/master/allocator/mesos/hierarchical.cpp
Lines 1425-1435 (original), 1426-1437 (patched)
<https://reviews.apache.org/r/53840/#comment248193>
As James commented, we should also stop the timer even if we return early but instead of adding it above `return Nothing()`, it seems appropriate to just use `metrics.allocation_run_interval.stop()` as the first statement of the method to mark the end of the batching delay.
src/tests/hierarchical_allocator_tests.cpp
Lines 3459-3463 (original), 3459-3462 (patched)
<https://reviews.apache.org/r/53840/#comment248194>
Why this change? Seems like the two ways of expressing the expectation are equivalent.
src/tests/hierarchical_allocator_tests.cpp
Lines 3537 (patched)
<https://reviews.apache.org/r/53840/#comment248196>
Seems inelegant to reuse the test for `allocation_run`, for `allocation_run` we are storing a full list of statistics in `statistics` but we are not doing the same for the new metric. Can we create an independent test?
src/tests/hierarchical_allocator_tests.cpp
Line 3544 (original), 3544 (patched)
<https://reviews.apache.org/r/53840/#comment248195>
Can it realistically be 0.0?
- Jiang Yan Xu
On April 21, 2017, 10:53 a.m., Anindya Sinha wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/53840/
> -----------------------------------------------------------
>
> (Updated April 21, 2017, 10:53 a.m.)
>
>
> Review request for mesos, James Peach and Jiang Yan Xu.
>
>
> Bugs: MESOS-6579
> https://issues.apache.org/jira/browse/MESOS-6579
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Added a metric 'allocator/mesos/allocation_run_interval_ms' which
> tracks the latency between scheduling an allocation to an actual
> allocation run.
>
>
> Diffs
> -----
>
> src/master/allocator/mesos/hierarchical.cpp 051f749dd5921a322ca930a042c31814616d38f9
> src/master/allocator/mesos/metrics.hpp ce486eb079dc44bf0bbfad8668426c7ae87c97b3
> src/master/allocator/mesos/metrics.cpp ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1
> src/tests/hierarchical_allocator_tests.cpp 33e7b455f8664858eb4f03727b076a10c80cd6e0
>
>
> Diff: https://reviews.apache.org/r/53840/diff/4/
>
>
> Testing
> -------
>
> All tests passed.
>
>
> Thanks,
>
> Anindya Sinha
>
>
Re: Review Request 53840: Metric in the allocator to track latency in
running allocations.
Posted by Anindya Sinha <an...@apple.com>.
> On May 15, 2017, 5:25 a.m., Jiang Yan Xu wrote:
> > src/tests/hierarchical_allocator_tests.cpp
> > Lines 3459-3463 (original), 3459-3462 (patched)
> > <https://reviews.apache.org/r/53840/diff/4/?file=1697516#file1697516line3459>
> >
> > Why this change? Seems like the two ways of expressing the expectation are equivalent.
Reinstated the original.
> On May 15, 2017, 5:25 a.m., Jiang Yan Xu wrote:
> > src/tests/hierarchical_allocator_tests.cpp
> > Lines 3537 (patched)
> > <https://reviews.apache.org/r/53840/diff/4/?file=1697516#file1697516line3539>
> >
> > Seems inelegant to reuse the test for `allocation_run`, for `allocation_run` we are storing a full list of statistics in `statistics` but we are not doing the same for the new metric. Can we create an independent test?
Created a separate test for `allocation_run_latency`.
- Anindya
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/53840/#review174871
-----------------------------------------------------------
On May 16, 2017, 6:23 a.m., Anindya Sinha wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/53840/
> -----------------------------------------------------------
>
> (Updated May 16, 2017, 6:23 a.m.)
>
>
> Review request for mesos, James Peach and Jiang Yan Xu.
>
>
> Bugs: MESOS-6579
> https://issues.apache.org/jira/browse/MESOS-6579
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Added a metric 'allocator/mesos/allocation_run_interval_ms' which
> tracks the latency between scheduling an allocation to an actual
> allocation run.
>
>
> Diffs
> -----
>
> src/master/allocator/mesos/hierarchical.cpp b75ed9a20a9a42f958cebbacd91e5e15b0d21394
> src/master/allocator/mesos/metrics.hpp ce486eb079dc44bf0bbfad8668426c7ae87c97b3
> src/master/allocator/mesos/metrics.cpp ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1
> src/tests/hierarchical_allocator_tests.cpp 08180b9975869de328f0c095dd3cddf0c84fbecf
>
>
> Diff: https://reviews.apache.org/r/53840/diff/5/
>
>
> Testing
> -------
>
> All tests passed.
>
>
> Thanks,
>
> Anindya Sinha
>
>