Private and public clouds, like AWS, Azure, or GCP, allow their users to deploy their high-availability (HA) workloads onto different Regions and Availability Zones (AZs) so that unexpected downtimes or physical outages in a data centre do not compromise their HA setups.
Designed as a safety mechanism for cloud users, Cloud-Enabled Livepatch is a feature to ensure that the patches are delivered to different regions or AZs based on a predefined schedule. In line with best practices around phased rollouts of updates, Cloud-Enabled Livepatch aims to minimise the risk of unexpected outages/transients due to wide-scale updates.
The schedule of delays, maintained by the Canonical Livepatch team, is carefully tailored to the needs of public cloud providers and their priorities over various regions/AZs. However, Cloud-Enabled Livepatch is not limited to public clouds. Private cloud owners who have deployed Livepatch On-prem to manage their fleet, can also enable this feature and configure it based on their own preferences.
The Cloud-Enabled Livepatch feature is enabled by default, and the users do not need further configuration. Normally, the Livepatch client extracts the host machine’s region/AZ information from the cloud-init configuration file, if any. However, based on their specific requirements, users have the option to override these settings, for which they can check this document for detailed instructions.
As an example, let us say an AWS user has deployed their workloads on two different AZs; us-east-1a and eu-central-1a. When a new patch is released, it will be first applied on the machines on one of the AZs, us-east-1a for example, and then after a delay of several hours, on the other AZ’s machines, eu-central-1a, will get the patch. However, this is just an example and the priority of AZs could be different in practice. The Livepatch team closely monitors the health status of the client machines, if there is an issue applying the patch during the initial rollout, the new patch will be blocklisted until the issue is resolved and other AZs will be kept safe from the same issue.