process-module-architecture_skill

Overview

This skill gives a concise architecture overview of the Process module that powers pipeline orchestration, scheduling, and execution in a CI/CD platform. It summarizes submodules, layered responsibilities, core data flows (pipeline CRUD, build lifecycle), and the event-driven execution path. Use it to orient development, debugging, or design changes to pipeline core logic.

How this skill works

The skill inspects the module split (api-process, biz-process, biz-base, biz-engine, model-process) and explains how requests travel from API entry through facade and engine services into the control layer. It maps the build lifecycle: triggers → interceptor chain → runtime service → event dispatcher → listeners → engine control (start → stage → container → task → end). It highlights DAO interactions and the RabbitMQ-based event dispatch used to decouple execution stages.

When to use it

• Developing or extending pipeline CRUD and template APIs
• Understanding or modifying build execution and scheduling logic
• Designing or refactoring engine control or task orchestration components
• Implementing event-driven listeners or message dispatch flows
• Diagnosing runtime issues in build start, stage, or task control paths

Best practices

• Read the overview first, then open the detailed layer document relevant to your change (API, service, engine, DAO, event).
• Keep facade services thin: coordinate validations, permissions, and model checks; push heavy orchestration to engine/biz-base services.
• Preserve the interceptor chain for global policies (locks, queueing, timer checks) to avoid duplicating controls across triggers.
• Use events to decouple synchronous request handling from long-running engine work; prefer durable messaging for build-critical events.
• Limit direct DB changes; prefer existing DAO services and jOOQ models to maintain schema consistency.

Example use cases

• Add a new webhook trigger that creates builds and flows through the interceptor chain into runtime service.
• Change concurrency policy: implement or modify a queue interceptor to alter parallel build limits.
• Extend an engine control (e.g., ContainerControl) to support new container lifecycle signals and agent startup events.
• Add a listener for custom build metrics that consumes PipelineBuildFinish events and forwards metrics to observability.
• Migrate part of pipeline state to a new table: update DAO, service layer, and ensure event listeners still find runtime state.