# Confluent Manager for Apache Flink Confluent Manager for Apache Flink® (CMF) is a central management component of Confluent Platform for Apache Flink that enables users to securely manage a fleet of Flink jobs across multiple [Environments](../configure/environments.md#cmf-environments). CMF supports two ways to create and manage Flink jobs; either as [applications](../jobs/applications/overview.md#cmf-applications) or [SQL statements](../jobs/sql-statements/overview.md#flink-sql-statements). This topic provides an overview of CMF and its architecture. CMF provides the following key features: - Job life-cycle management for Flink jobs. - Integration with Confluent Platform for authentication and authorization (RBAC). - Well-defined REST APIs and command-line interfaces (CLIs). - Management of Flink workloads across multiple Kubernetes clusters from a single CMF instance. For more information, see [Manage Multiple Kubernetes Clusters in Confluent Manager for Apache Flink](../configure/kubernetes-clusters.md#cmf-kubernetes-clusters). The following image shows the high-level architecture of how CMF fits in.  CMF sits next to other Confluent Platform components such as Confluent Server and Connect. Metadata for CMF, such as existing Flink environments, is stored in a database: an embedded SQLite database by default, or an external PostgreSQL, Microsoft SQL Server, or IBM Db2 database. For more information, see [Configure Storage for Confluent Manager for Apache Flink](../installation/storage.md#cmf-storage). Most interactions with CMF occur through its [REST API](../clients-api/rest.md#af-rest-api). You can use these APIs directly for maximum flexibility, and the same APIs are exposed with the Confluent CLI. They are also integrated with the Confluent for Kubernetes (CFK) for a native Kubernetes experience. CMF also serves the [CMF UI](../installation/ui.md#cmf-ui-config) and an optional [MCP server](../clients-api/mcp.md#cmf-mcp) for AI agents. #### IMPORTANT Do not manually modify Flink deployment resources: CMF manages Flink deployment resources internally and you should not modify them manually. Manual modifications to Flink deployment resources can cause conflicts and unexpected behavior. ## Get started To get started with CMF, see the following topics: - [Get Started with Applications](../get-started/get-started-application.md#cpf-get-started) - [Get Started with FlinkSQL](../get-started/get-started-statement.md#get-started-flinksql) - [Install CMF with Helm](../installation/helm.md#install-cmf-helm) ## Related content - [Confluent CLI Flink Operations](https://docs.confluent.io/confluent-cli/current/command-reference/flink/index.html) - [REST APIs for Confluent Manager for Apache Flink](../clients-api/rest.md#af-rest-api) - [Manage Multiple Kubernetes Clusters in Confluent Manager for Apache Flink](../configure/kubernetes-clusters.md#cmf-kubernetes-clusters) - [Manage Flink Using Confluent for Kubernetes](https://docs.confluent.io/operator/current/co-flink-overview.html)