TDengine

What it is

TDengine is a purpose-built time-series database written in C, designed for IoT, industrial telemetry, and monitoring workloads. It uses a "one table per device/sensor" data model, ships with a SQL-like query interface, built-in stream processing, and native data subscription (similar to a lightweight Kafka bolted onto the storage engine). It's positioned to replace the common stack of InfluxDB + Kafka + Spark for time-series-heavy pipelines by folding ingestion, storage, caching, and stream compute into a single binary. Category placement under "App Definition and Development" is a CNCF taxonomy artifact — in practice this is a database, not a dev framework.

Who builds it and why

Maintained primarily by TDengine (the company, formerly TAOS Data), a commercial open-core vendor. 317 contributors on GitHub is a healthy number for a database project, but expect the bulk of commits to come from paid engineers, not a broad volunteer community — that's typical for open-core time-series databases (see also InfluxDB, QuestDB). CNCF maturity level: not publicly available, and TDengine is not a CNCF project as far as public listings show — it's independently governed by the vendor.

Production readiness signal

24,943 GitHub stars and a commit as recent as 2026-06-30 indicate active, ongoing development — this isn't an abandoned or coasting project. Latest release ver-3.4.1.6 suggests a mature versioning cadence past the 3.x rewrite (TDengine did a significant architecture overhaul between 2.x and 3.x, so if you're evaluating this, don't rely on old 2.x benchmarks or docs). No public data on founding date, but the 3.x line and contributor count imply a project with several years of production hardening behind it. AGPL-3.0 licensing is a signal worth flagging on its own (see Pricing below) — it affects how you can legally embed or offer this as a service.

Who should use this

• Teams ingesting high-volume, high-cardinality time-series data from sensors, devices, or industrial equipment where write throughput and storage compression matter more than ad hoc query flexibility.
• Organizations currently running InfluxDB + Kafka + a stream processor who want to consolidate into fewer moving parts and reduce operational surface area.
• Engineers comfortable with SQL who want a lower-code alternative to building custom stream processing on top of a general-purpose database.

Who should NOT use this

• Teams needing general-purpose relational or document storage — this is not a substitute for Postgres, MongoDB, or similar; forcing non-time-series data into it will fight the data model.
• Companies that need to offer TDengine as a hosted/managed service to third parties — AGPL-3.0 has strong copyleft implications for network-exposed services; get legal review before building a SaaS offering on top of it without a commercial license from the vendor.
• Shops without in-house time-series or observability expertise — the "one table per device" model and stream processing DSL have a learning curve; if you just need dashboards and basic metrics, a hosted Prometheus/Grafana stack is less operational overhead.
• Anyone requiring a large, vendor-neutral open-source community for long-term risk mitigation — this is closer to single-vendor open-core than a broad multi-stakeholder project.

Alternatives

• InfluxDB — the incumbent time-series database, larger ecosystem and community, but historically weaker on high-cardinality data and clustering (varies by version/license tier).
• QuestDB — Postgres-wire-compatible time-series database, strong for SQL-native teams wanting minimal new tooling.
• TimescaleDB — a Postgres extension approach; best fit if you want time-series capability without leaving the Postgres ecosystem.

Pricing

Core TDengine is open source under AGPL-3.0 — free to self-host and use. The vendor offers a commercial/enterprise edition (TDengine Enterprise) with additional features (e.g., enhanced high availability, security, support) under a paid license — exact pricing is not publicly available and requires contacting the vendor. If you plan to expose TDengine as a network service to third parties, confirm licensing obligations under AGPL before shipping.