RethinkDB

What it is

RethinkDB is an open-source, distributed document database built in C++, best known for changefeeds — a native push-query mechanism that lets applications subscribe to real-time updates on query results instead of polling. It stores JSON documents, shards and replicates natively across a cluster, and exposes a query language (ReQL) rather than SQL. It was purpose-built for apps that need live data (dashboards, collaborative tools, multiplayer backends) without bolting a pub/sub layer onto a traditional database.

Who builds it and why

RethinkDB started as a VC-funded company that shut down commercial operations in 2016 after failing to find a sustainable business model in a crowded database market (competing against MongoDB, and later against Firebase for the real-time use case). The company donated the project to the Linux Foundation, and it's since been maintained by a volunteer community. The 247 contributors and steady commit history (latest activity through March 2026) indicate the project didn't die with the company — but it's community-driven, not vendor-backed, and the pace and priorities reflect that: stability and compatibility fixes more than new feature velocity.

Production readiness signal

Mixed signal, and worth reading carefully. On one hand: 26,999 GitHub stars, 247 contributors, and commits as recent as March 2026 show the project is alive, not abandoned. Latest release v2.4.4 suggests a mature, incrementally-patched codebase rather than active reinvention.

On the other hand: the license field shows NOASSERTION — meaning GitHub's license detection can't confirm what license actually governs the repo. For any production or commercial use, this is not a footnote, it's a blocker until you manually verify the licensing terms in the repo yourself (historically RethinkDB used Apache 2.0, but confirm current state before you build on it). CNCF maturity level is not publicly available, which suggests it isn't formally in the CNCF's sandbox/incubating/graduated pipeline despite being categorized under App Definition and Development — treat it as a community open-source project, not a CNCF-governed one.

Who should use this

• Teams building real-time features (live dashboards, collaborative editing, presence/notifications) who want changefeeds as a first-class primitive instead of layering Kafka/Redis pub-sub on top of a document store.
• Engineers who already run RethinkDB in production and are maintaining/extending existing systems — the upgrade path and community are active enough to support that.
• Hobbyists and small teams prototyping real-time apps where operational simplicity matters more than ecosystem size.

Who should NOT use this

• Anyone starting a greenfield project in 2026 without existing RethinkDB expertise. The talent pool, tooling ecosystem, and managed-hosting options have shrunk since 2016; you'll be operating this yourself.
• Teams needing strong ecosystem integration (Terraform providers, cloud-managed offerings, mature ORMs across languages) — MongoDB, PostgreSQL, and Firebase all have deeper tooling.
• Anyone who needs SQL compatibility or an existing team fluent in SQL — ReQL is powerful but it's another language to learn with a shrinking base of shared knowledge.
• Regulated or enterprise environments where the ambiguous license status (NOASSERTION) is a legal or procurement blocker — get this resolved before you touch it in a commercial product.
• Teams that need vendor support contracts or SLAs — there is no commercial entity behind RethinkDB anymore.

Alternatives

• MongoDB — similar document model, vastly larger ecosystem, mature change streams for real-time use cases, commercial support available.
• Firebase Realtime Database / Firestore — managed, real-time-first, less operational burden, but vendor lock-in to Google Cloud.
• PostgreSQL with LISTEN/NOTIFY or logical replication — if you don't need document flexibility, get real-time-ish behavior on a battle-tested relational engine with a much larger talent pool.

Pricing

Fully open source. No commercial tier, no managed offering, no vendor support contract exists post-2016. You run it, you patch it, you support it — factor that into any adoption decision.