9 months ago
38
We’re excited to announce a new partnership with Elastic to support their Elastic Cloud on Kubernetes (ECK) product on GKE in Autopilot mode of operation.
Customers can now take advantage of the combined benefits of GKE Autopilot and ECK for their search, observability and security needs, all running on infrastructure that provides a fully managed Kubernetes experience, dynamic just-in time scalability, and a low total cost of ownership (TCO).
What is the GKE mode of operation?
GKE Autopilot is the default and recommended mode of operation to run your applications on GKE. Autopilot fully manages your cluster for you, but still gives you access to the full Kubernetes API, and all the control you need to fine-tune your workload’s performance.
Autopilot manages your nodes and capacity for you by adding and removing nodes from your cluster in response to your workload’s changing capacity demands and characteristics, helping provide a strong experience. Simply specify the resource requests and your scale and infrastructure requirements in your podspec (for example, generic compute, CPU intense, or GPU acceleration) and GKE provisions the right infrastructure just-in-time to meet your workloads needs.
GKE Autopilot also manages your Kubernetes upgrades, helps provide a secure posture out-of-the-box, and helps us stand by our reliability commitment with one of the first-ever Pod-level service level agreements (SLA), in addition to the control-plane SLA. And since you only pay for the resources your workloads use, you don’t have to worry about wasted cluster capacity, improving costs. Because of that, and the other benefits mentioned above, Autopilot provides you with one of the best TCOs in the industry.
What is ECK?
Elastic Cloud on Kubernetes, or ECK for short, is the official Elastic Operator for Kubernetes.
ECK is a great way to run the Elastic Stack on your cloud-native Kubernetes environment, as it encompasses key operational knowledge and best practices, and streamlines the operations and use of the Elastic Stack.
In addition to automating operational and cluster administration tasks, ECK helps streamline the entire Elastic-Stack-on-Kubernetes experience. It makes upgrading, resizing clusters, and configuring various advanced functionality like cross-cluster search and replication, index lifecycle management or index snapshot policies and repositories as simple as adding a few lines of code to a resource definition.
With ECK, you can easily manage one, tens, or hundreds of Elasticsearch clusters at large scale.
Why ECK on GKE Autopilot?
With Elastic Cloud on Kubernetes running on GKE Autopilot, you gain enhanced operational efficiencies, as it removes the need to know, plan and size Kubernetes node pools to match the size of your Elastic nodes.
You have several options for where to run your Elastic Stack: using self-managed solutions like ECK on GKE Autopilot, or Elastic’s Elasticsearch Service, which is Elastic Cloud’s hosted and fully managed option.
You may want to choose ECK, for example, if you have complex networking requirements with multiple network domains and tight controls over which network domains your data traverses through. In such cases, a managed solution like Elastic Cloud might be less suitable for you. However, ECK still provides many important operational advantages, especially when paired with a fully automated Kubernetes platform such as GKE Autopilot.
For example, one customer was using Elasticsearch to provide functional search services on their product catalog and technical search services to monitor their API usage.
Consumption of these services was seasonal, but because the Elasticsearch clusters were deployed on self-managed infrastructure, they had long provisioning cycles and couldn’t easily adapt to seasonal usage.
Additionally, the customer had to deploy across several providers and regions to achieve global distribution. As such, they decided to use Kubernetes, which allowed them to apply the same standards, architecture, and tools across all their deployments and their global footprint.
To standardize the technical stack as much as possible, they needed a managed service for Kubernetes, either GKE Standard or GKE Autopilot. They opted for GKE Autopilot, whose billing model and reduced total cost of ownership (TCO), let them focus on their business rather than their infrastructure.
Migrating to GKE Autopilot also provided them better operational efficiencies, with its just-in-time global scalability, while ECK provided the operational efficiencies of a managed Elastic Stack.
Next steps — get started with GKE Autopilot and ECK
Together, the teams at GKE and Elastic are excited about this partnership, which will help joint customers serve customers across the globe with managed services. To get up and running with Autopilot quickly, check out the following resources: