A self-contained Python script that uses the grafanalib library to generate a Grafana dashboard JSON for visualizing Bull queue metrics exposed by the exporter. It demonstrates a clean, parameterized approach to dashboard-as-code and can be reused with minimal adaptation for any Prometheus exporter that exposes Bull metrics with the default metric prefix.

The core class managing Bull queues, metric registration, and auto-discovery. It reuses a single Redis client for Bull's 'client' connection while creating separate connections for subscribers/bclient, which reduces resource usage. The `discoverAll` method uses async iteration over ioredis's scanStream to incrementally discover queues by matching Redis key patterns, a modern, memory-efficient pattern. The class is well-organized with clear separation of concerns (discovery, update, completion event listening) and handles cleanup gracefully. Although tied to Bull's API, the design patterns (shared client, event-based metrics, async iteration) are transferable to other queue-based metric collectors.