Ocean by Spot provides key functions that deliver a serverless infrastructure experience for containerized applications by taking care of auto scaling, right sizing and pricing optimization. These features enable users to focus on their applications while the infrastructure is automatically provisioned and managed by Ocean.
To further enhance the serverless experience, Ocean also addresses the often manual process of re-provisioning compute infrastructure. Re-provisioning typically occurs when there is a change on the application layer, for example when a new app version requires a bigger root disk or there is a change to the user data script. With Ocean, users benefit from a “roll” action that replaces all the instances in the cluster. Scheduled in advance by the user, once triggered, the process is performed gradually and the container status is monitored.
Initially, the cluster roll only re-provisioned instances on an all-or-nothing basis, replacing all the nodes in the cluster at once. Today, we are happy to share that we have added granular control for Ocean’s cluster roll to support a “per launch specification roll”, giving users the ability to perform the roll even on specific instances if they require. This update enables a faster, more efficient process to replace underlying infrastructure as Ocean users only need to roll the relevant part of the infrastructure, and monitor only the instances that are being replaced.
The new ability is expected to soon be integrated in the full cluster roll process as an automatic failover mechanism, and in the Spot console for easy activation in the launch specification level.
For more details and to activate this feature, please refer to the relevant Spot documentation to learn more.