The current setting of ceilometer notification workers can result in
an excessive number in multi-node system with high core controllers.
It should be limited to 10 max to avoid unnecessary system resources
consumption and to improve vcpu scaling responsiveness. The degraded
vcpu scaling performance is caused by time taken to rebuild the
transformer caches in all workers following a swact/service restart
as well as the increased cpu sample age in these caches.
Tests conducted: sanity, nightly sanity and 25 iterations of vcpu
scaling tests in multi-node system with 32-core controllers.
Closes-Bug: #1790679
Change-Id: I47b7f14a75846fd9765e9a2ea2fb3ca3a6cad01a
Signed-off-by: Tee Ngo <email address hidden>
Reviewed: https:/ /review. openstack. org/599737 /git.openstack. org/cgit/ openstack/ stx-config/ commit/ ?id=86d60c3c298 d5737acff61f3a1 04731c00ac7fb2
Committed: https:/
Submitter: Zuul
Branch: master
commit 86d60c3c298d573 7acff61f3a10473 1c00ac7fb2
Author: Tee Ngo <email address hidden>
Date: Tue Sep 4 14:06:26 2018 -0400
Limit number of notification workers to 10 max
The current setting of ceilometer notification workers can result in
an excessive number in multi-node system with high core controllers.
It should be limited to 10 max to avoid unnecessary system resources
consumption and to improve vcpu scaling responsiveness. The degraded
vcpu scaling performance is caused by time taken to rebuild the
transformer caches in all workers following a swact/service restart
as well as the increased cpu sample age in these caches.
Tests conducted: sanity, nightly sanity and 25 iterations of vcpu
scaling tests in multi-node system with 32-core controllers.
Closes-Bug: #1790679
Change-Id: I47b7f14a75846f d9765e9a2ea2fb3 ca3a6cad01a
Signed-off-by: Tee Ngo <email address hidden>