Amazon Timestream
Serverless time-series database for IoT, metrics, and operational telemetry
TDengine
High‑performance, cloud‑native time‑series database for IoT scale
TDengine delivers fast ingestion, compression, and AI‑enhanced analytics for billions of IoT sensors, supporting Linux, macOS and Kubernetes deployments with built‑in stream processing and RESTful access.
Overview
Overview
TDengine is designed for developers, data engineers, and data scientists who need to ingest, store, and analyze massive IoT or industrial sensor streams. It delivers sub‑second ingestion rates, columnar compression, and a simple SQL‑like query language while exposing built‑in AI services for forecasting, anomaly detection, and classification.
Deployment & Operations
The database runs natively on Linux and macOS and can be packaged as containers or deployed via Helm charts on Kubernetes, supporting both public and private cloud environments. Its distributed architecture separates compute from storage, uses RAFT for consensus, and includes tools such as taosBenchmark, taosdump, and a RESTful taosAdapter, reducing the need for additional middleware.
Key capabilities include automatic sharding, time‑based partitioning, integrated stream processing, and an AI agent (TDgpt) that connects to external foundation models. This makes TDengine a single‑stack solution for real‑time analytics and long‑term archival of petabyte‑scale time‑series data.
Highlights
High‑performance ingestion and columnar compression
Built‑in AI agent (TDgpt) for forecasting and anomaly detection
Cloud‑native distributed architecture with RAFT and Kubernetes support
Integrated stream processing, caching, and RESTful API
Pros
- Scales to billions of data points per day
- Simplifies pipelines with native tools and AI services
- Strong open‑source community and extensive connector ecosystem
- Separation of compute and storage enables flexible deployment
Considerations
- Windows support limited to enterprise edition
- Build requires recent GCC and CMake versions
- Cross‑compilation not currently provided
- Advanced features may need extra configuration
Managed products teams compare with
When teams consider TDengine, these hosted platforms usually appear on the same shortlist.
Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.
Fit guide
Great for
- Large‑scale IoT sensor fleets requiring real‑time analytics
- Teams that want integrated AI forecasting without external services
- Kubernetes or hybrid‑cloud environments seeking a native TSDB
- Organizations preferring a single solution over multiple components
Not ideal when
- Small projects with minimal data volume
- Environments that need native Windows OSS support
- Users requiring out‑of‑the‑box graphical management tools
- Workloads that depend on traditional relational database semantics
How teams use it
Real‑time fleet telemetry processing
Ingests millions of vehicle metrics per minute, runs on‑edge compute, and provides predictive alerts via TDgpt.
Industrial sensor data archiving
Stores petabytes of time‑series data with high compression, enabling fast historical queries for maintenance analysis.
Smart city environmental monitoring
Collects air‑quality and traffic sensor streams, applies built‑in stream processing for aggregation, and serves dashboards through REST API.
Predictive maintenance for manufacturing equipment
Uses AI agent to forecast failures, reducing downtime by automatically triggering maintenance tickets.
Tech snapshot
C83%
C++11%
Shell2%
CMake1%
Go1%
Yacc1%
Frequently asked questions
What operating systems does TDengine support?
Linux and macOS are supported for the open‑source edition; Windows is available only in the enterprise edition.
How does TDengine handle high cardinality?
It uses a native sharding and compression scheme that stores billions of series efficiently, avoiding the performance penalties typical of other TSDBs.
Can TDengine be deployed on Kubernetes?
Yes, it includes native distributed design, RAFT consensus, and Helm charts for seamless Kubernetes deployment.
What AI capabilities are built in?
The integrated TDgpt agent can connect to foundation models for time‑series forecasting, anomaly detection, imputation, and classification.
Is there a RESTful interface?
Yes, the taosAdapter provides a REST API that lets any application interact with the database without native drivers.
Project at a glance
Active- Stars
- 24,681
- Watchers
- 24,681
- Forks
- 4,988
LicenseAGPL-3.0
Repo age6 years old
Last commityesterday
Primary languageC
Last synced yesterday