Architecture Overview

Technical architecture of AutoDoctor across agent, persistence, API service, and dashboard layers, including runtime boundaries and data flow.
Categories:
Tags:

Who This Is For

  • Developers extending diagnostics/remediation modules.
  • Operators debugging interactions between agent and API.

High-Level Components

  • agent/: PowerShell orchestration, modules, telemetry, SQLite writes
  • server/api/: FastAPI app, SQLite query layer, service wrapper
  • server/dashboard/: static dashboard + Chart.js polling client
  • config/: INI template and runtime config
  • installer/: build and Inno packaging scripts

Execution Flow

  1. Agent entrypoint (AutoDoctor.ps1) initializes paths and DB.
  2. Module engine executes registered modules in sequence.
  3. Root cause and remediation run, then results are persisted.
  4. Telemetry JSON and latest_run.json are written.
  5. API reads SQLite + metadata and serves endpoints.
  6. Dashboard polls API every 5 seconds.

Service Runtime Flow

  • Service wrapper (autodoctor_service.py) resolves runtime paths and mode.
  • Child API process starts from bundled binary or source script depending on mode.
  • API binds host/port based on precedence (registry -> INI -> env -> defaults).

Key Design Characteristics

  • Local-first storage and runtime
  • SQLite as shared state between writer (agent) and reader (API)
  • Configurable service registration mode for constrained environments
  • Clear separation of collection (agent) and presentation (API/dashboard)

Next Steps