Udhay Ravindran
Share
In June 2022, we launched the DigitalOcean Kubernetes HA control plane with 99.95% uptime SLA. Since then, you were able to create DigitalOcean Kubernetes (DOKS) clusters with the new control plane and add High Availability (HA) when provisioning. Today, we are excited to announce that you can migrate any DOKS cluster to the new control plane and enable HA your favorite way.
We can’t wait for you to try the much more redundant, fault-tolerant, and resilient new control plane. Among many other enhancements, the new control plane offers automatic healing and recovery, dynamic resource allocation, and quick feature updates.
The new control plane automatically heals and recovers any of its unhealthy components, making it more resilient to unexpected failures. On-demand CPU and memory allocation let DigitalOcean Kubernetes dynamically adapt to variable usage patterns. The new control plane architecture makes it easy to ship updates to the control plane continuously.
Eliminate the single point of failure (SPOF) that exists in Kubernetes by adding High Availability to your DigitalOcean Kubernetes cluster. High Availability replicates control plane components to ensure resilient clusters, increase fault tolerance, and protect against control plane outages. We offer a 99.95% uptime SLA for High Availability for all HA-enabled clusters. Check out the DigitalOcean Kubernetes HA SLA for more details and the credit-claim process.
After upgrading your Kubernetes clusters to v.1.22.X or higher, you can enable High Availability in the UI or programmatically. Add HA to your existing clusters programmatically with doctl v.1.86.0 or later using the flag --ha=true. To turn on High Availability in the UI, navigate to the control plane card in the Cluster Overview tab and follow the instructions. Be aware that you can’t disable HA once it’s enabled on a cluster. For more details, see the enable HA docs.
You might already be running DigitalOcean Kubernetes clusters on the new control plane and can enable HA anytime. If you need to migrate, follow our guide to eliminate any disruptions. To make sure, check your Kubernetes clusters in your cloud control panel. You’ll see a message like the following if your clusters can upgrade
To check if your cluster is running on a HA control plane see your Cluster Overview tab it has a Control Plane card with the HA status along with the guidance to enable it.
If you’d like to upgrade to enable High Availability only DigitalOcean Kubernetes version equal to 1.22.X or greater can enable it and unlock future developments. Kubernetes clusters older than v1.21.X can update as usual, in order to jump to the next minor version you must be running the latest patch version 1.21.14-do.0. The first doctl version to officially support HA enablement will be 1.87.0.
When migrating your clusters to the upgraded control plane you may experience up to two minutes of disruption to your data plane network connection where you can’t contact the Cluster API. You can add HA once after migrating to the new control plane. Follow this migration guide to minimize any impact to your workloads.
Migrations to the new control plane will also be enabled through the required upgrades going forward after the initial phase of voluntary migrations. Please look out for an extra note about the migration capability in the regular email notifications about required upgrades.
The new control plane is available in all DigitalOcean regions and is completely free because we don’t charge for the control plane. We only charge for the underlying usage of products by the cluster like nodes and Load Balancers. High Availability pricing remains the same at $40 per month for each enabled cluster. Our pricing page has more information.
At DigitalOcean, we’re committed to partnering in your Kubernetes journey, accelerating your growth, and scaling your business. We look forward to helping grow your business on DigitalOcean Kubernetes. Reach out to our Solutions Experts to get more information or for help migrating to DigitalOcean Kubernetes.
Happy Coding,
Udhay Ravindran,Senior Product Manager, Kubernetes
Share
Bikram Gupta, Ingo Gottwald, Piyush Srivastava, Braden Bassingthwaite, Udhay Ravindran