Today we’re happy to announce the release of Elastic Cloud on Kubernetes (ECK) 1.1.0. This is our first minor release of ECK since general availability in January, and it builds on our vision to make operating the Elastic Stack on Kubernetes as easy as possible.
ECK 1.1.0 includes enhanced support for configuring cross-cluster features and introduces declarative users and roles.
Enhanced cross-cluster support
ECK gives users the ability to deploy and manage multiple Elasticsearch clusters with ease. This simplicity also allows users to follow best practices by splitting separate teams, tenants, or use cases into their own smaller Elasticsearch clusters. Users who operate global environments may need to replicate their data across geographies to ensure the lowest latency for end users performing searches.
ECK 1.1.0 introduces support for remote clusters, which makes setting up cross-cluster search or cross-cluster replication within the same Kubernetes environment just a few lines of configuration. Users leveraging Elasticsearch 7.6 and above with ECK can also use cross-cluster search and replication across multiple global Kubernetes environments.
Some use cases include:
- Observability, where you may have relevant logs, metrics, and APM data spread out into different Elasticsearch deployments
- Search across multiple Elasticsearch versions simultaneously, in case you have sets of data that aren’t ready to be upgraded to a new Elasticsearch version
- Search across multiple geographically dispersed data centers, allowing you to store data locally and search globally
To use remote clusters in a single Kubernetes environment, simply provide the name and namespace of the remote cluster, and ECK will automatically handle the configuration for you. This includes setting up all authentication and authorization between the clusters, adding the necessary certificate authorities to the trusted list, and managing their rotation if necessary.
Declarative users and roles
ECK 1.1.0 also introduces support for declarative users and roles. ECK handles creating the user and role definitions based on the Kubernetes Secrets provided, eliminating the need to call the Elasticsearch API to create them. With ECK you can specify your users, credentials, and roles for your Elasticsearch clusters as code. This allows you to perform GitOps-style deployments and even plug into existing CI/CD pipelines.
This is especially useful for those who want to maintain multiple identical environments, such as separate development and production Kubernetes clusters — or multiple Kubernetes clusters in multiple geographic locations. You can check your configuration into version control and let ECK handle the orchestration.
Plus even more improvements
This release also includes myriad quality-of-life improvements, including:
- Additional validation of resources at admission time: This prevents a broad class of errors such as incorrect indentation and provides faster feedback loops when creating or modifying Elastic resources.
- Automatic Kibana encryption key specification: This allows you to create multiple Kibana instances and maintain session tokens no matter which backend instance users access.
- Role-based access control for associations: For more secure environments, ECK now allows you to specify a service account to use when associating a Kibana or APM resource with a backend Elasticsearch resource. This allows for more granular control over what associations are permitted.