Skip to content

Instance High Availability service (OpenStack Masakari)#

This article describes operations with Instance High Availability (Instance HA).

The Instance High Availability service (OpenStack Masakari) allows cloud users to ensure that their instances are automatically evacuated from a failed hypervisor. It provides several types of monitoring services:

  • Instance monitor — checks the liveness of instance processes.
  • Introspective instance monitor — improves instance high availability within OpenStack environments by monitoring and identifying system-level failures through the QEMU Guest Agent.
  • Host monitor — checks the liveness of compute hosts and runs as part of the Node Controller in Rockoon.

The introspective instance monitor is disabled by default and must be explicitly enabled in the Masakari configuration.

Enabling the Instance HA service#

To enable the Instance HA service your need to add instance-ha to the service list in OpenStackDeployment custom resource:

spec:
  features:
    services:
      - instance-ha

Enabling introspective instance monitor#

To enable the introspective instance monitor in the Masakari service, update the spec:features:masakari:monitors:introspective section in OpenStackDeployment custom resource:

spec:
  features:
    masakari:
      monitors:
        introspective:
          enabled: true