Self-Managed Connectors on Confluent Cloud
Confluent Cloud supports a hybrid model for connectors. You can run your own connectors in a self-managed Connect cluster and still get unified visibility into them from Confluent Cloud. Unified Stream Manager (USM) extends this visibility to self-managed connectors that produce to and consume from a Confluent Cloud Kafka cluster.
This lets you run a connector wherever your data lives, such as inside your own network next to an on-premises or otherwise private data source or sink, while the Kafka cluster it produces to and consumes from remains a fully managed Confluent Cloud cluster.
This capability applies only when the Connect cluster’s metadata cluster is Confluent Cloud Kafka. If your Connect cluster instead produces to and consumes from a self-managed Confluent Platform Kafka cluster, see Confluent Platform Kafka as the Metadata Cluster and Monitor Confluent Platform Connectors with USM instead.
New to USM? Start with the Unified Stream Manager overview to learn what USM is and how it connects your self-managed environment to Confluent Cloud.
What you can do
USM gives you monitoring visibility into self-managed connectors. You can monitor your self-managed connectors from the same Connectors page that lists your Confluent Cloud fully managed connectors, including connector status, throughput, and per-connector health and performance details. For monitoring details, see Monitor Self-Managed Connectors Running Against Confluent Cloud Kafka.
Before you begin
Ensure that your Confluent Cloud environment has the Advanced Governance package, which USM requires.
Deploy the USM Agent in your self-managed environment. The Agent collects connector metadata and metrics and sends them to Confluent Cloud over a secure network connection. A registered Connect cluster shows no connector data until the Agent is running. For setup instructions, see Deploy the USM Agent.
Register a self-managed Connect cluster against a Confluent Cloud Kafka cluster so that USM can begin collecting its connector metadata and metrics. For registration steps, see Confluent Cloud Kafka as the Metadata Cluster.
Limitations
For the full list of limitations, see Monitor Self-Managed Connectors Running Against Confluent Cloud Kafka.