// the find
unitedrhino/things
联犀是基于golang,gozero框架构建的SaaS云原生微服务物联网平台(iot)，支持MQTT、CoAP、HTTP、TCP、阿里云、腾讯云、涂鸦云,物美等云云对接及设备协议，实现海量多样设备连接。支持多租户、多应用和项目管理，内置小程序功能,适应多样化业务场景.支持单体、集群和微服务部署，适应不同企业规模和需求。支持mysql和pgsql，满足数据存储需求,适用于智能家居、智慧楼宇...
unitedrhino/things is a Go IoT platform built on go-zero, targeting SaaS deployments where you need multi-tenant device management, protocol bridging (MQTT/CoAP/HTTP/TCP), and time-series telemetry storage. The notable design choice is a single codebase that can run as a monolith on 2GB RAM or scale out to full microservices — same code, different config. Aimed at Chinese enterprises building smart-building, industrial, or agricultural IoT products.
- The monolith-to-microservices deployment switch via config (not code) is genuinely well-executed — go-zero's RPC abstraction makes this work, and it lets small teams start cheap and scale without a rewrite.
- Protocol and cloud bridge coverage is real work, not just claims: they handle Alibaba Cloud IoT, Tencent Cloud IoT, Tuya, and Modbus alongside standard MQTT/TCP, which saves weeks of integration grunt work.
- TDengine for time-series telemetry is the right call — querying device history at IoT scale on a relational DB always ends badly, and they've made the right infrastructure choice here.
- The hierarchical thing-model (product category → product → device schema) is well thought out and maps cleanly to how real IoT deployments are structured.
- The README is pure SEO keyword stuffing — '物联网' appears 200+ times in obvious repetition. This reflects poorly on the team's judgment and the English docs are sparse, meaning non-Chinese teams will hit walls fast.
- The full stack requires MySQL/PG + Redis + NATS + TDengine + EMQX + etcd + MinIO running simultaneously — seven infrastructure services before your first device connects. The 'min' Docker configs exist but aren't well documented.
- Private key files committed to the repo (doc/godLei.pem, doc/godLei.ppk) — even if these are example/expired keys, it signals careless security hygiene in the docs/tooling layer.
- 627 stars for a project claiming enterprise-grade, million-device scale is a small community footprint, meaning limited external bug reports, plugins, or third-party integrations — you'll mostly be on your own when something breaks.