Deploy and Manage Confluent Platform Using Confluent for Kubernetes
Confluent for Kubernetes (CFK) is a cloud-native management control plane for deploying and managing Confluent Platform in your Kubernetes private cloud environment. It offers a standard and simple interface to customize, deploy, and manage Confluent Platform through a declarative API.
CFK is a Kubernetes Deployment whose lifecycle is managed by Helm, but all the Confluent Platform components define their own custom resource definitions (CRDs) that you can manage like any other Kubernetes resource.
CFK actively monitors the custom resources (CRs) to ensure their state matches the desired state.
Note
CFK is the next generation of Confluent Operator. For Confluent Operator 1.x documentation, see Confluent Operator, or use the version picker to browse to a specific version of the documentation.
For an overview of CFK, see the Confluent for Kubernetes Offers Cloud-Native Kafka Automation blog.
The following video introduces CFK.
What you can do with CFK
With CFK, you can:
Deploy declaratively: Configure and deploy Apache Kafka®, Connect workers, ksqlDB, Schema Registry, Confluent Control Center, Confluent REST Proxy, and KRaft or ZooKeeper, along with application resources such as topics and rolebindings, as code. Customize components with server properties, Java Virtual Machine (JVM), Log4j, and Log4j 2 configuration overrides.
Automate upgrades: Roll out configuration changes and version upgrades with automated rolling updates that avoid impact to Kafka availability.
Scale on demand: Scale components and storage with single commands and built-in reliability checks.
Stay resilient: Restore a Kafka pod with the same broker ID, configuration, and storage after a failure, and spread partition replicas across racks or zones with automated rack awareness.
Schedule workloads: Use Kubernetes labels and annotations to identify and organize workloads, and use tolerations and pod or node affinity to control where workloads run and how resources are used.
Secure by default: Apply granular role-based access control (RBAC), authentication, and TLS network encryption, use auto-generated certificates, and integrate credential management systems such as HashiCorp Vault.
Monitor: Aggregate metrics with Java Management Extensions (JMX) and Jolokia, and export aggregated metrics to Prometheus.
What CFK does not do
CFK deploys Confluent Platform on Kubernetes only. To install Confluent Platform on bare-metal hosts or virtual machines, use the manual or Ansible Playbooks installation instead.
CFK manages Confluent components and Confluent application resources, not arbitrary third-party Kubernetes workloads.
CFK manages self-managed Confluent Platform in your own Kubernetes environment. It does not manage Confluent Cloud.
New to CFK?
If you are getting started, use the following as a suggested path.
Learn the architecture and core concepts behind Confluent for Kubernetes.
Get familiar with the key terms you see throughout the docs.
Deploy Confluent Platform on Kubernetes in a few steps.
Review supported environments and options for production.