Re: Review Request 53840: Metric in the allocator to track latency between allocation cycles.
> On April 20, 2017, 6:04 p.m., James Peach wrote: > > If I understand this correctly, this is publishing a time series of the > > idle time between allocator runs. Is that correct? If so, what insight does > > this metric give and how should operators interpret it? Why is a simple > > counter of allocator runs not sufficient? > > Anindya Sinha wrote: > So as to determine how frequently (or infrequently) allocation cycles are > being run by the allocator. The counters provide the current count of > allocation runs. As we discussed with @xujyan, this is intended to measure the queueing delay for allocation dispatches. I'm still not convinced that this is a helpful metric. As implemented, this appears to measure the time between allocation intervals. I guess you would start the timer in `HierarchicalAllocatorProcess::allocate` and stop it in `HierarchicalAllocatorProcess::_allocate`. - James --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53840/#review172526 --- On April 19, 2017, 8:51 p.m., Anindya Sinha wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/53840/ > --- > > (Updated April 19, 2017, 8:51 p.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 consecutive allocation cycles. > > > Diffs > - > > src/master/allocator/mesos/hierarchical.cpp > 0059ccead90f32491591990c791e7fa905a864b7 > src/master/allocator/mesos/metrics.hpp > ce486eb079dc44bf0bbfad8668426c7ae87c97b3 > src/master/allocator/mesos/metrics.cpp > ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1 > src/tests/hierarchical_allocator_tests.cpp > cdf1f15b7802439b28405ca8f6634ce83e886630 > > > Diff: https://reviews.apache.org/r/53840/diff/3/ > > > Testing > --- > > All tests passed. > > > Thanks, > > Anindya Sinha > >
Re: Review Request 53840: Metric in the allocator to track latency between allocation cycles.
> On April 20, 2017, 6:04 p.m., James Peach wrote: > > If I understand this correctly, this is publishing a time series of the > > idle time between allocator runs. Is that correct? If so, what insight does > > this metric give and how should operators interpret it? Why is a simple > > counter of allocator runs not sufficient? So as to determine how frequently (or infrequently) allocation cycles are being run by the allocator. The counters provide the current count of allocation runs. - Anindya --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53840/#review172526 --- On April 19, 2017, 8:51 p.m., Anindya Sinha wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/53840/ > --- > > (Updated April 19, 2017, 8:51 p.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 consecutive allocation cycles. > > > Diffs > - > > src/master/allocator/mesos/hierarchical.cpp > 0059ccead90f32491591990c791e7fa905a864b7 > src/master/allocator/mesos/metrics.hpp > ce486eb079dc44bf0bbfad8668426c7ae87c97b3 > src/master/allocator/mesos/metrics.cpp > ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1 > src/tests/hierarchical_allocator_tests.cpp > cdf1f15b7802439b28405ca8f6634ce83e886630 > > > Diff: https://reviews.apache.org/r/53840/diff/3/ > > > Testing > --- > > All tests passed. > > > Thanks, > > Anindya Sinha > >
Re: Review Request 53840: Metric in the allocator to track latency between allocation cycles.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53840/#review172526 --- If I understand this correctly, this is publishing a time series of the idle time between allocator runs. Is that correct? If so, what insight does this metric give and how should operators interpret it? Why is a simple counter of allocator runs not sufficient? - James Peach On April 19, 2017, 8:51 p.m., Anindya Sinha wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/53840/ > --- > > (Updated April 19, 2017, 8:51 p.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 consecutive allocation cycles. > > > Diffs > - > > src/master/allocator/mesos/hierarchical.cpp > 0059ccead90f32491591990c791e7fa905a864b7 > src/master/allocator/mesos/metrics.hpp > ce486eb079dc44bf0bbfad8668426c7ae87c97b3 > src/master/allocator/mesos/metrics.cpp > ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1 > src/tests/hierarchical_allocator_tests.cpp > cdf1f15b7802439b28405ca8f6634ce83e886630 > > > Diff: https://reviews.apache.org/r/53840/diff/3/ > > > Testing > --- > > All tests passed. > > > Thanks, > > Anindya Sinha > >
Re: Review Request 53840: Metric in the allocator to track latency between allocation cycles.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53840/ --- (Updated March 7, 2017, 11:41 p.m.) Review request for mesos and Jiang Yan Xu. Changes --- Changed the metric to track latency between allocation cycles, as opposed to tracking the last allocation cycle. Summary (updated) - Metric in the allocator to track latency between allocation cycles. Bugs: MESOS-6579 https://issues.apache.org/jira/browse/MESOS-6579 Repository: mesos Description (updated) --- Added a metric 'allocator/mesos/allocation_run_interval_ms' which tracks the latency between consecutive allocation cycles. Diffs (updated) - src/master/allocator/mesos/hierarchical.cpp 0059ccead90f32491591990c791e7fa905a864b7 src/master/allocator/mesos/metrics.hpp ce486eb079dc44bf0bbfad8668426c7ae87c97b3 src/master/allocator/mesos/metrics.cpp ec987fe0ce8605bf9e7c0ccbe3ac384deaf53be1 src/tests/hierarchical_allocator_tests.cpp cdf1f15b7802439b28405ca8f6634ce83e886630 Diff: https://reviews.apache.org/r/53840/diff/3/ Changes: https://reviews.apache.org/r/53840/diff/2-3/ Testing --- All tests passed. Thanks, Anindya Sinha
