Hazelcast loves Kubernetes. Thanks to the dedicated Hazelcast Kubernetes plugin, you can use dynamic auto-discovery. Hazelcast on Kubernetes can also run in multiple topologies: embedded, client-server, or as a sidecar. What’s more, thanks to the Helm package manager and the dedicated Hazelcast Helm Chart, you can deploy a fully functional Hazelcast server in literary minutes. I already described it in the Hazelcast Helm Chart blog post, which covered the scenario when the client and the server were both deployed in the same Kubernetes cluster.
In this blog post, let’s focus on a more difficult scenario, where you’d like to set up your own on-premises Hazelcast on a Kubernetes cluster and then use it with a client located outside that cluster.