First public release of the Distributed Task Queue Engine (dtq).

What it is: a Redis Streams-backed distributed task queue in Python/FastAPI/asyncio.

• Consumer-group workers with cooperative multitasking and bounded concurrency
• Retries with exponential backoff and delayed scheduling; dead-letter queue for poison tasks
• Idempotent execution via idempotency keys; at-least-once delivery, honestly documented
• Backpressure: queue depth limits with HTTP 429 responses
• Real-time observability: WebSocket event stream, Prometheus metrics, React dashboard
• dtq CLI for submitting, inspecting, retrying, and managing tasks

Verification: 169 tests passing (unit, integration, concurrency, e2e), ruff and mypy clean, demo script 6/6 green.

Measured, not claimed: ~1.1k tasks/sec ingestion, ~212 tasks/sec completion (3 workers), p50 latency ~1.8s on a local run.