The biggest change is Couchbase’s vastly increased automation capabilities, which will be of particular interest to customers with large Kubernetes deployments.
Today, NoSQL database provider Couchbase has announced Version 2.0 of the Couchbase Autonomous Operator for Kubernetes. This version of Autonomous Operator considerably extends the relatively vanilla capabilities of 1.0. Its new automation capabilities are particularly relevant to larger deployments. It adds more capabilities of the core Couchbase platform, which were absent from the initial release. It now lets enterprises run a Couchbase Data Platform cluster on more Kubernetes services, addressing concerns about AWS lock-in. It also provides the customer with greater visibility of and control over Autonomous Operator, which will be extended further this summer when Couchbase Cloud is scheduled to become available.
Autonomous Operator for Kubernetes is a packaging of the Couchbase server that includes running inside of Docker and overseeing Kubernetes operators. It’s a specialized install, which is a core part of the Couchbase platform. The 1.0 version hit General Availability in August 2018.
“The focus of Version 1.0 was making sure the deployment exercise works very well,” said Jeff Morris, Couchbase’s VP of Product and Solutions Marketing. “As it evolved, we started gathering more requirements from customers about what they wanted to see.”
Those requests began with bringing more core Couchbase functionality into Autonomous Operator.
“They wanted Autonomous Operator for Kubernetes to be able to take advantage of more Couchbase features, such as cross-database replication, so you are not bound to any one geographic location,” Morris said. “That’s a big regular feature, but automating that for Autonomous Operator raised issues of how the database will rebalance itself as it replicates across different areas. Data centre replication is something customers wanted earlier, as was fully autonomous and self-healing backup.”
Major changes to these capabilities were facilitated by Kubernetes modifying their design so developers could add custom resources.
“This allowed us, in 2.0, to adjust the way Autonomous Operator works to accommodate that, to create an automatic kind of system, tackle the operationalization of using Couchbase and get into more granular levels with Autonomous Operator,” Morris indicated. “Now, for example, if a node goes away and loses connectivity, Autonomous Operator will be able to recognize it and help heal that particular instance.”
The increased automation capability, which includes Automated Security Management, Automated Backup/Restore Management and Automated Cross Datacenter Replication Management, is critical for users with larger Kubernetes deployments, because it automates the deployment and management of hundreds of clusters across multiple setups, regions, private and public clouds. It’s something that is critical for larger users – but those are also the Couchbase users who are working with Autonomous Operator today.
“Between 15-20% of our install base is using Autonomous Operator,” Morris indicated. “It doesn’t matter as much if you have a 3-node cluster. It does matter more for larger customers.”
Version 2.0 now supports a much broader range of cloud providers, including Red Hat OpenShift, Google Kubernetes Engine GKE, Amazon Elastic Kubernetes Service EKS, and Microsoft’s Azure Kubernetes Service AKS.
“Most customers are now worried about getting tied too much to one particular cloud provider, specifically AWS,” Morris said. “They also want services in Azure and GCP and now we support that.”
Version 2.0 is improved at supporting health stats for clusters, integrated with the cloud-native Prometheus monitoring system. It also provides users with control and visibility, including a fine-grained advanced Kubernetes Operator Security Model, and certificate management using Mutual TLS Support.
Control will be further extended this summer, with the scheduled release of Couchbase Cloud, which is also built around Autonomous Operator.
“It’s our managed database as-a-service offering,” Morris stated. “It’s in beta now and will ship in the summer. It’s automated, and facilitated by Autonomous Operator capabilities. Instead of us owning and provisioning on behalf of our customers, we install it in the customer’s infrastructure, so they can take operation without it ever leaving their network. They have a higher degree of control. Our approach is run in the customer’s private cloud accounts, and that’s enabled by Autonomous Operator.”
Morris stressed that these capabilities are ahead of the broad market.
“To us, this is one of the deepest and most comprehensive kinds of using Kubernetes,” he said. “We are only seeing a handful of other vendors do similar things. This deploying software inside the customer’s account with autonomous operation like you get from traditional SaaS, facilitated by Autonomous Operator, will also be very interesting.”