Back at the Victoria PTG the Octavia team discussed deprecating the spares pool capability of the amphora driver[1]. This would follow the standard OpenStack deprecation process[2]. There are a number of reasons this was proposed: 1. It adds a lot of complexity to the code. 2. It can't be used with Active/Standby load balancers due to server group (anti-affinity) limitations in Nova. 3. It provides only 15-30 seconds of speedup when provisioning a new load balancer on production clouds. 4. It makes supporting Octavia availability zones awkward as we have to boot spares instances in each AZ. 5. It can be confusing for people when it is enabled as there are always extra amphora running and being automatically recreated. Due to these reasons a patch has been proposed to deprecate spares pool support in the amphora driver: https://review.opendev.org/741686 Please comment on that patch and/or join the weekly Octavia IRC meeting if you have any concerns with this deprecation plan. Michael [1] https://etherpad.opendev.org/p/octavia-virtual-V-ptg [2] https://governance.openstack.org/tc/reference/tags/assert_follows-standard-d...