How to Connect Kafka to PostgreSQL in Real Time
Connecting Kafka to PostgreSQL in real time means streaming events from Kafka topics into PostgreSQL-compatible tables for querying. The three main approaches: Kafka Connect JDBC Sink (traditional), Debezium + reverse CDC (complex), and RisingWave (simplest — Kafka source + SQL + PostgreSQL protocol). RisingWave acts as both the Kafka consumer and the queryable database.
Three Approaches Compared
| Approach | Components | Latency | SQL Processing | Complexity |
| Kafka Connect JDBC | Kafka + Connect + PostgreSQL | Seconds | ❌ (direct sink) | Medium |
| Flink + PostgreSQL | Kafka + Flink + PostgreSQL | Sub-second | ✅ (Flink SQL) | High |
| RisingWave | Kafka + RisingWave | Sub-second | ✅ (PG SQL) | Low |
RisingWave Approach (Simplest)
-- Step 1: Create Kafka source
CREATE SOURCE kafka_events (user_id INT, event VARCHAR, amount DECIMAL, ts TIMESTAMP)
WITH (connector='kafka', topic='events', properties.bootstrap.server='kafka:9092')
FORMAT PLAIN ENCODE JSON;
-- Step 2: Create materialized view (processes + stores)
CREATE MATERIALIZED VIEW event_summary AS
SELECT event, COUNT(*) as cnt, SUM(amount) as total
FROM kafka_events GROUP BY event;
-- Step 3: Query with any PostgreSQL client
-- psql, DBeaver, Python psycopg2, etc.
SELECT * FROM event_summary;
No JDBC connector, no Flink, no separate PostgreSQL. RisingWave IS the PostgreSQL-compatible database that consumes Kafka.
Frequently Asked Questions
Do I need Kafka Connect to get Kafka data into PostgreSQL?
Not if you use RisingWave. RisingWave natively consumes Kafka topics and exposes results via PostgreSQL protocol. No Kafka Connect, no JDBC sink, no middleware.
Can I transform Kafka data before loading into PostgreSQL?
With RisingWave, yes — SQL transformations (joins, aggregations, filters) are applied as materialized views. With Kafka Connect JDBC, transformations require SMTs (Single Message Transforms) which are limited.
