Tango Flow: A Low‑Code Workflow Orchestration Platform for Cloud Music Backend

Background and demands : Cloud Music’s front‑back collaboration relies on RESTful APIs and RPC‑driven data assembly in the controller layer. This creates heavy contract dependencies, high communication cost, and poor reusability of BFF (Backend‑For‑Frontend) interfaces. GraphQL‑based BFFs were introduced in 2022, but they still suffer from limited orchestration capability, resource waste, and unclear operational responsibilities.

Goal : Build an assembly‑style architecture that integrates Cloud Music’s server‑side tech stack, provides basic logic orchestration, and exposes the results via gateway APIs or unified SDKs. The platform is positioned as a key component of a full‑chain low‑code initiative for backend development.

Orchestration concept : A complete workflow can chain resources in sequence, branch, loop, or parallel. Two resource categories are supported – business‑domain services (RPC, HTTP, FaaS) and tool‑domain services (middleware, AI tools, platform services). By assembling these assets, flexible service work‑flows are created.

Product architecture :

Logical orchestration state: Tango Flow aggregates existing OX assets (business RPC/HTTP interfaces) and allows users to build, test, and publish workflows.

Logical runtime state: The Flow Engine can be embedded in the gateway, BFF, or business services, effectively replacing BFF capabilities.

Key features :

Self‑developed orchestration engine : Supports static workflow runtime, can be exposed as a gateway API or integrated via SDK.

Custom DSL protocol : Defines meta‑information (triggers, RPC nodes, Groovy scripts) and logical flow; future AI‑assisted prompts are planned.

Cluster hosting mechanism : Multi‑tenant hosted clusters reduce resource waste; clusters can be shared across applications and scaled on demand.

Environment isolation : API gateway handles routing; Tango Flow engine manages environment‑specific routing.

Visual workflow builder : Drag‑and‑drop UI with component area, logic canvas, property panel, version history, and DSL view.

Parameter selection : Nodes can consume upstream results directly or via JsonPath expressions.

Debugging : Full‑workflow and node‑level debugging (RPC, HTTP, Groovy) without publishing; supports environment selection and records.

Mock mechanism : Request and response mocks for RPC, HTTP, and Groovy to accelerate parallel development.

Continuous release : Full‑lifecycle release pipeline with multi‑branch parallel development, environment‑specific strategies, rollback, and approval checkpoints.

Monitoring & alerting : RT, error‑rate, and access‑distribution metrics with role‑based alerting.

Application scenarios :

BFF scenario – solves orchestration, stability, and cost issues.

Backend business scenario – provides a unified SDK to reuse platform‑asset services, reducing development cost and accelerating delivery.

Future roadmap :

Server‑side low‑code evolution – more trigger types and component integration.

Front‑back low‑code unification – model‑driven full‑stack low‑code.

Platform experience & efficiency optimization – based on user feedback.

Conclusion : Tango Flow addresses the core orchestration challenges of Cloud Music’s backend, enables complex workflow composition, improves stability and resource efficiency, and serves as a foundation for a low‑code, full‑chain development paradigm.