How-to guides

Each guide solves one problem in 30–150 lines. Code first, gotchas at the end.

Production setup

• Configure retries — attempts, backoff types, jitter, queue-wide vs per-job.
• Idempotent add — addUnique, XADD ... IDMPAUTO, the at-least-once safety boundary.
• Enable result storage — opt in to storeResults and poll with waitForResult.

Reliability

• Route to the DLQ — UnrecoverableError, every DlqReason, when each fires.
• Replay the DLQ — chasqui dlq replay and the shim helper. Idempotency caveats.
• Schedule repeatable jobs — cron, every, MissedFiresPolicy.

Operations

• Observe the engine — MetricsSink, Prometheus adapter, chasqui events.

Performance

• Tune for throughput — concurrency, batched acks, payload size, the enableAutoPipelining lesson.

Deployment

• Produce from AWS Lambda — TLS / rediss://, ElastiCache + IAM auth, module-scope clients, the operational gotchas (NAT 350s, ENI cap, IAM token rotation).

Migration

• Migrate from BullMQ — API mapping, what’s compat, what’s intentionally different.
• Migrate from Sidekiq or Celery — conceptual mapping for ruby / python shops.