nats-best-practices

Choosing between Core NATS and JetStream depends on your use case: Core NATS offers simple, fire-and-forget messaging, while JetStream provides persistence, exactly-once delivery, and fault tolerance for robust message handling. This skill offers guidance on making this decision based on scalability and system integration requirements.

Scaling workers consuming from the same subject in JetStream is best achieved using Pull Consumers, which allow multiple workers to pull messages concurrently. This skill provides best practices for configuring Pull Consumers to balance load and ensure high throughput without message loss.

Exactly-once delivery and idempotency in NATS JetStream are implemented using robust publisher patterns with deduplication strategies. This skill demonstrates how to configure Streams and Consumers with appropriate retention policies and idempotent publishing to ensure exactly-once processing.

Setting up a Dead Letter Queue (DLQ) in JetStream is essential for handling messages that fail processing after retries. This skill covers DLQ implementations and considerations, ensuring fault tolerance by isolating problematic messages without blocking the main queue.

Handling graceful shutdown in Go with NATS requires properly draining connections and acknowledging processed messages before exiting. This skill details best practices for graceful shutdown to prevent message loss during system redeploys or unexpected terminations.

Using this Skill for NATS message routing and system design patterns requires a Go-first approach and existing knowledge of NATS and message-oriented architecture. It focuses on advanced decision-making and load balancing, making it unsuitable for users without foundational Go and NATS experience.