From f10c58c3437d8ec25d70a41782dbe8d68c8b2024 Mon Sep 17 00:00:00 2001 From: Ilya Sokolov Date: Sun, 13 Sep 2015 21:58:51 +0300 Subject: [PATCH] [feature request] Check queue perfdata Change-Id: Iedf407fe5c9f4ac4da39156ce324de734f2e4440 --- doc/feature_request/check_queue_perfdata.rst | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) create mode 100644 doc/feature_request/check_queue_perfdata.rst diff --git a/doc/feature_request/check_queue_perfdata.rst b/doc/feature_request/check_queue_perfdata.rst new file mode 100644 index 0000000000..c09ff5a1d5 --- /dev/null +++ b/doc/feature_request/check_queue_perfdata.rst @@ -0,0 +1,19 @@ +=========================== +Check queue perfdata +=========================== + + +Use case +———— +Sometimes OpenStack services use common messaging system very prodigally. For example neutron metering agent sending all database table data on new object creation i.e https://review.openstack.org/#/c/143672/. It cause to neutron degradation and other obvious problems. +It will be nice to have a way to track messages count and messages size in queue during tests/benchmarks. + + +Problem description +————————— +Heavy usage of queue isn’t checked. + + +Possible solution +———————— +* Before running tests/benchmarks start process which will connect to queue topics and measure messages count, size and other data which we need.