Data systems

opendata-oss/

opendata

Rust databases for logs, metrics, vectors and write buffering, all built on SlateDB so the data lives in object storage.

What’s new here

OpenData picks one storage engine, SlateDB, an LSM tree that runs on object storage, and builds several databases on top of it. Each database only has to define its own data layout and query semantics. The pitch is that the whole fleet then shares one operational profile and one set of tooling.

What it does

OpenData is a collection of four databases today:

  • Buffer: write-path buffering and ingestion pipelines.
  • Log: replayable event streams and durable data movement.
  • Timeseries: Prometheus-compatible metrics storage.
  • Vector: large-scale semantic search and retrieval.

SlateDB handles write batching, tiered caching and compaction. It gets snapshot isolation from atomic manifest updates using S3’s compare-and-set. OpenData itself is meant to supply the shared layer above storage: service infrastructure, a service catalog, admin tooling, distributed state and testing frameworks. The README lists most of that as roadmap, not shipped.

Who it’s for

Engineers who run databases in the cloud and want durability and cost to come from object storage. It also suits people who want to follow or contribute to an early project. The README says plainly that it is early and building in the open.

Try it

The README has no quick start at the top level. Each database (buffer, log, timeseries, vector) has its own README, which the top-level README says documents that database’s roadmap. The latest release is opendata-vector/v0.2.2. It ships prebuilt vector archives for Apple Silicon macOS, Linux (aarch64 and x86_64) and Windows (x86_64), so you do not need to build that one from source.

To read the code:

git clone https://github.com/opendata-oss/opendata

How mature is it

Created in November 2025 and now at 541 stars, 48 forks and 18 contributors, under the MIT license. There are 28 releases, and the latest, opendata-vector/v0.2.2, is dated 2026-07-14. The last push was 2026-08-07 and the API shows no commits in the last 90 days as of 2026-10-26. 56 issues and 14 pull requests are open. The README calls the project early, and the shared service layer, admin tooling and distributed mode are still unchecked roadmap items.