Elastic follows users on the K8s train, proclaiming ECK production ready

Elastic follows users on the K8s train, proclaiming ECK production ready

Elastic Cloud on Kubernetes (ECK) has made it through beta phase and is now available in version 1.0 which doesn’t add a lot of exciting stuff, but does make the tool more reliable.

Elastic introduced ECK in May 2019, in order to provide “a seamless way of deploying, managing, and operating the Elastic Stack on Kubernetes”. The first major, and therefore production ready, release contains functions to deploy and manage Elasticsearch clusters and Kibana, TLS certificate management, as well as helpers to scale and upgrade systems. And to make it even more en vogue, advanced topologies “such as dedicated master and machine learning nodes” are meant to be supported as well.

Since the last update, the tool has learned to automatically add certificates for a webhook’s Secret and ValidatingWebhookConfiguration objects. Thanks to multi-namespace cache support, users can now restrict an operator to a small set of namespaces. The ECK team was also able to reduce downtime during pod recycling, and fitted the init containers with default requests and limits.

Apart from that, it has put some work into figuring out license handling, since – as usual with Elastic products – the core is free to use while more advanced features require a subscription. This model led to some contention last year, when AWS launched its own Elasticsearch distro claiming that the mixture of proprietary and free to use code in the project’s codebase meant a “lack of clarity as to what customers who care about open source are getting and what they can depend on”.

ECK users that have been trying the project but didn’t look into the 1.0 beta, might encounter breakage, since nodes and nodeCount have been renamed to nodeSets and count amongst other things. A look into the beta’s release notes might help to find solutions if you don’t want to start from scratch anyway. 

For the future, Elastic’s principal product manager Anurag Gupta promised the addition of a dedicated UI and autoscaling capabilities amongst other things. More details can be found in his blog post celebrating the release.