At a glance
| Sources | MySQL, PostgreSQL, SQL Server, MariaDB, Oracle, MongoDB, Amazon DynamoDB, Google Cloud Spanner |
| Targets | Apache Kafka topics on Confluent Cloud or Confluent Platform (then any Confluent sink connector) |
| CDC mechanism | Log-based via Debezium (binlog, logical decoding, SQL Server CDC); Oracle LogMiner or XStream Out |
| Initial load | Snapshot modes including initial, initial_only, when_needed and no_data; incremental snapshots |
| Schema evolution | Schema history topic and Schema Registry formats (Avro, Protobuf, JSON Schema) |
| Latency | Not documented as a figure |
| Conflict handling | Not applicable: one-way capture into Kafka |
| Monitoring | Connector status and metrics in Confluent Cloud; offset management API |
| Deployment | Fully managed on Confluent Cloud (AWS, Azure, Google Cloud) or self-managed Kafka Connect |
| Licensing | Per task-hour (CDC: contact Confluent) plus USD 0.025 per GB; premium Oracle connectors need a subscription |
Choose Confluent's CDC connectors when Kafka on Confluent Cloud is already your event backbone and you want Debezium-style capture without running Kafka Connect yourself. They are a feature of the Confluent platform rather than a standalone tool: CDC task-hour prices are quote-only, and you still pay for the Kafka cluster and data transfer.
Overview
Confluent offers a family of change data capture source connectors for Kafka Connect. On Confluent Cloud the fully managed list includes MySQL, PostgreSQL and Microsoft SQL Server CDC Source V2 (Debezium), MariaDB CDC Source, MongoDB CDC Source (Debezium), Google Cloud Spanner CDC Source (Debezium), Amazon DynamoDB CDC Source, Oracle CDC Source and Oracle XStream CDC Source. Each writes row-level inserts, updates and deletes to Kafka topics, by default one topic per table, in Avro, Protobuf, JSON Schema or plain JSON.
The Debezium-based connectors package the open source Debezium engine as a managed service: for example, the MySQL CDC Source V2 connector supports MySQL 8.0 and 8.4, snapshot modes such as initial, no_data and when_needed, incremental snapshots through signals and a single task. Confluent's own Oracle CDC Source reads redo logs through Oracle LogMiner and supports Oracle 19c, 21c and 23ai Enterprise Edition, while Oracle XStream CDC Source uses Oracle's XStream Out interface. Self-managed versions run on Confluent Platform or any Kafka Connect cluster.
Key features
- Fully managed CDC connectors on Confluent Cloud, with no Kafka Connect workers to operate
- Debezium-based V2 connectors for MySQL, PostgreSQL and SQL Server, plus MongoDB and Spanner
- Confluent premium Oracle CDC Source (LogMiner) and Oracle XStream CDC Source connectors
- Output in Avro, Protobuf or JSON Schema with Schema Registry, or schemaless JSON
- Initial and incremental snapshots, and offset management through the Confluent Cloud API
- Private networking, AWS IAM authentication for RDS and Aurora, and client-side field-level encryption options
- Single message transforms, including custom SMTs at an extra task-hour charge
Integrations and ecosystem
Pros and cons
Pros
- Managed Debezium-based capture without running Kafka Connect yourself
- Covers the main relational sources plus MongoDB, DynamoDB and Spanner
- Commercial Oracle CDC options with vendor support
- Tight integration with Schema Registry and the wider Confluent connector catalogue
Cons
- Task-hour prices for the CDC connectors are not published ("Contact us")
- Only delivers to Kafka; other targets need a separate sink connector
- Some connectors are limited to one task, and the V1 Debezium connectors reached end of life in March 2026
- Oracle connectors do not work with Oracle Autonomous Database
Pricing
Fully managed connectors on Confluent Cloud are billed per task-hour plus USD 0.025 per GB of data transfer (measured before compression), with prices varying by region. Confluent's Connect pricing page shows "Contact us" for the task-hour price of the MySQL, PostgreSQL and SQL Server CDC Source V2 (Debezium), MariaDB CDC, Oracle CDC Premium and Oracle XStream CDC Premium connectors, so CDC task pricing is by quote. You also pay for the Kafka cluster. The self-managed Oracle CDC Source connector runs free for a 30-day trial, then needs a Confluent connector subscription.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Teams already standardised on Confluent Cloud or Confluent Platform that want supported, managed database CDC into Kafka topics.
Typical use cases
- Streaming OLTP changes from MySQL, PostgreSQL or SQL Server into Kafka for event-driven services
- Capturing Oracle changes into Kafka as part of moving off batch extracts
- Feeding real-time data into warehouses or lakes through Confluent sink connectors
- Building a single governed stream of database changes with Schema Registry
Who should look elsewhere
- Teams without Kafka who want changes delivered straight to a warehouse: see Estuary or Fivetran.
- Teams that want free, self-run CDC: use open source Debezium directly.
Verdict
Confluent's CDC connectors are a practical way to get Debezium and Oracle capture into Kafka without operating Connect clusters, backed by a commercial vendor now owned by IBM. They make most sense when Confluent is already your streaming platform; the quote-only task pricing and Kafka-only output make them less attractive as a standalone CDC tool.
Frequently asked questions
Are Confluent's CDC connectors based on Debezium?
Several are. The MySQL, PostgreSQL and SQL Server CDC Source V2 connectors, MongoDB CDC Source and Spanner CDC Source are labelled Debezium. The Oracle CDC Source and Oracle XStream CDC Source connectors are Confluent's own.
How are fully managed CDC connectors priced?
Confluent Cloud bills connectors per task-hour plus USD 0.025 per GB of data transfer. For the CDC connectors the task-hour price shows "Contact us", so ask Confluent for a quote, and add the cost of the Kafka cluster.
What happened to the V1 Debezium connectors on Confluent Cloud?
Confluent deprecated the V1 MySQL, PostgreSQL and SQL Server CDC Source (Debezium) connectors on 9 January 2025 and they reached end of life on 31 March 2026. Existing ones keep running but cannot be created; migrate to the V2 connectors.
Is Confluent now part of IBM?
Yes. IBM completed its acquisition of Confluent on 17 March 2026. Confluent said it would continue as a distinct brand within IBM.
Sources
- Confluent Cloud connector overview
- Confluent Cloud Connect pricing
- MySQL CDC Source V2 (Debezium) connector for Confluent Cloud
- Oracle XStream CDC Source connector for Confluent Cloud
- Oracle CDC Source connector (self-managed)
- IBM completes acquisition of Confluent (17 March 2026)
Checked November 2026.
Listed in: Change Data Capture & Replication.
How we research reviews and what star ratings mean: our editorial method.