What problem does it solve? Long-running operations block HTTP requests, tightly coupled services fail together, and background jobs get lost without retries. This Skill provides decision frameworks and working patterns for building event-driven architectures, background job processing, and service decoupling with message brokers. ## Core Features & Use Cases - Broker Selection Guidance: Decision tree comparing Kafka (event streaming), RabbitMQ (complex routing), NATS (request-reply), Redis Streams (simple queues), and Temporal (workflow orchestration) with throughput and latency benchmarks. - Production Patterns: Event naming conventions, schema structure, dead letter queues, idempotency for exactly-once processing, and anti-patterns to avoid. - Multi-Language Examples: Working code for Celery (Python), BullMQ (TypeScript), Kafka producers/consumers, and Temporal saga workflows, plus frontend integration via SSE for job status updates. - Use Case: Offload a 5-minute report generation from your API by enqueueing a Celery task, then stream progress updates to a React component via Server-Sent Events. ## Quick Start Ask the agent to set up a background job queue with Celery and Redis for processing image uploads asynchronously, including retry logic and a dead letter queue.
