Auto Scaling & Sentinel

Franza Cavalcante will discuss how to use Autoscaling with Monasca and a project he implemented (using Heat) called Sentinel.

Sentinel is a tool to place a VM on each hypervisor and demonstrate that the hypervisor is working properly with respect to scheduling, networking, and that storage is working properly. (Some other similar implementations would call this a canary instance.) Charter Communications uses Sentinels in their OpenStack cloud as an "uptime" metric.

Monasca (Monitoring at Scale) provides metrics that can be used by Heat to trigger scale up and scale down.

Franza has previously given a presentation to this Meetup on Heat.