• Automatic Queue Processing — The system checks for pending orders every few seconds. When orders exist, it processes them one by one without delay. When the queue is empty, it falls back to a periodic check — no manual intervention needed.
• Self-Regulating Speed — Processing adapts to workload automatically. Busy queues are handled as fast as possible; empty queues consume no resources. The system scales its own pace without configuration.
• Built-In Control Panel — A web interface lets you start, stop, and monitor processing in real time. See remaining orders with a progress bar, check the current state, and reset the queue — all from one screen.
• Queue Status & Management — Queue state is available as an API, so external systems can query progress or trigger resets programmatically alongside the built-in UI.

• Polling Reaktor Registration — registerPollingReaktor creates a timer-based Reaktor with a configurable interval (5 seconds) and autoStart: false. The platform manages the execution lifecycle — start, stop, and timing are handled via built-in REST endpoints.
• Conditional Flow with aktorIfElse — The core logic is a single aktorIfElse: if orders exist (aktorHasOrders), process one (aktorProcessOrder); otherwise return false to signal idle. The boolean return value directly controls whether the next execution is immediate or delayed.
• Sync and Async Aktors — aktorHasOrders is a synchronous createAktorFunction (simple array check), while aktorProcessOrder is an async createAktorFunctionAsync (simulates processing time). Both patterns coexist in the same graph.
• Helper Reaktors as API Endpoints — registerReaktorDefinition creates two additional endpoints: one for queue status queries, one for resetting the queue. These are plain Reaktors that return data — no polling, no timer — used by the UI for real-time monitoring.
• Integrated Custom UI — The public/ directory ships a control panel built with plain JavaScript and Bootstrap. It uses window.getOperaideAuthToken() for authentication and auto-refreshes every two seconds when the processor is running.

The Polling Coffee Queue app registers three endpoints. The polling Reaktor is the core: its createReaktor returns an aktorIfElse that checks for orders and either processes one (returning true for immediate re-execution) or signals idle (false for interval wait). Two helper Reaktors provide queue status and reset functionality as standard REST endpoints. The custom UI in public/ ties them together — calling the polling control endpoints (start/stop/status) and the helper Reaktors (queue status/reset) to provide a real-time dashboard. The entire app is a single TypeScript file plus the UI assets.