RisingWave

What is RisingWave?

RisingWave is an open source streaming database that processes continuous event streams with standard SQL. You define materialized views over your incoming data, and RisingWave keeps their results up to date incrementally as new events arrive — then serves those results to applications over the PostgreSQL wire protocol. It is written in Rust and uses object storage (like S3) for state, so compute and storage scale independently.

What is RisingWave best for?

Teams building real-time features — live dashboards, monitoring, fraud detection, personalization, feature pipelines — who want fresh answers over streaming data without running a separate stream processor and a serving database. It fits best when your team already knows SQL and Postgres and wants to avoid the operational weight of a Flink cluster.

What can RisingWave do?

• Ingest from Kafka, Pulsar, Kinesis, database CDC (Postgres/MySQL), and webhooks
• Keep incrementally-maintained materialized views fresh with sub-second latency
• Serve query results directly over the PostgreSQL wire protocol — no separate serving store
• Decouple compute from storage using S3-compatible object storage for state
• Sink results out to Kafka, databases, and Apache Iceberg tables
• Scale horizontally across nodes and run on Docker, Kubernetes (Helm), or an operator

Is RisingWave free?

Yes — the RisingWave database is open source under the Apache 2.0 license and free to self-host; you only pay for the servers and object storage you run it on. RisingWave Cloud is the paid managed option, billed on RisingWave Units (RWUs) starting at $0.227 per RWU-hour on the Basic plan, with storage charged separately per GB-month.

Where does RisingWave fall short?

• It is a streaming database, not a general-purpose OLTP or data-warehouse engine — it is built to maintain streaming materialized views, not to run heavy ad-hoc analytical scans over historical data the way Snowflake or a batch warehouse does.
• It is Postgres wire-compatible, but not a drop-in Postgres replacement — SQL coverage, functions, and extensions differ, so complex existing queries may need rewriting.
• Its architecture leans on S3-style object storage for state, which adds read/write latency and cost characteristics you have to plan around; getting the best performance still requires tuning memory and cluster sizing.

What does RisingWave replace?

RisingWave replaces the traditional real-time stack of Kafka plus a stream processor plus a serving database with one system. In practice it stands in for parts of a Confluent Cloud streaming setup and can take over real-time workloads that teams would otherwise push into Snowflake or Databricks — keeping results continuously fresh instead of re-running batch jobs. Its closest open source comparison is Apache Flink, which is a stream processor rather than a database and cannot store and serve results on its own.

FAQ

Is RisingWave open source? Yes. The core database is licensed under Apache 2.0 — a permissive OSI-approved license — and the source is on GitHub. RisingWave Cloud is a separate paid managed service.

Can I self-host RisingWave for free? Yes. Self-hosting the Apache-2.0 database is free; you pay only for your own compute and object storage. RisingWave Cloud is the paid, hosted alternative.

Is RisingWave a good Flink alternative? For SQL-first teams, often yes — it uses PostgreSQL-compatible SQL, stores and serves results as a database, and needs far less tuning than a Flink cluster. Flink remains more flexible for low-level, code-heavy stream processing.

What do I need to run RisingWave? A server or Kubernetes cluster and S3-compatible object storage for state. It runs via Docker Compose for a single-node start, or Helm / an operator on Kubernetes for a scaled, production deployment.