iii-dead-letter-queues（iii 死信队列）

Dead Letter Queues

Comparable to: SQS DLQ, RabbitMQ dead-letter exchanges

Key Concepts

Use the concepts below when they fit the task. Not every queue failure needs manual DLQ intervention.

• Jobs move to a DLQ after exhausting max_retries with exponential backoff (backoff_ms * 2^attempt)
• Each DLQ entry preserves the original payload, last error, timestamp, and job metadata
• Redrive via the built-in iii::queue::redrive function or the iii trigger CLI command
• Redriving resets attempt counters to zero, giving jobs a fresh retry cycle
• Always investigate and deploy fixes before redriving — blindly redriving repeats failures
• DLQ support available on Builtin and RabbitMQ adapters

Architecture

A queue consumer fails processing a job. The engine retries with exponential backoff up to max_retries. Once exhausted, the message moves to the DLQ. An operator inspects the failure, deploys a fix, then redrives the DLQ to replay all failed jobs.

iii Primitives Used

Reference Implementation

See ../references/dead-letter-queues.js for the full working example — inspecting DLQ status,

Also available in Python: ../references/dead-letter-queues.py

Also available in Rust: ../references/dead-letter-queues.rs redriving failed jobs via SDK and CLI, and configuring retry behavior.

Common Patterns

Code using this pattern commonly includes, when relevant:

• await iii.trigger({function_id: 'iii::queue::redrive', payload: {queue: 'payment'}}) — redrive via SDK
• iii trigger --function-id='iii::queue::redrive' --payload='{"queue": "payment"}' — redrive via CLI
• iii trigger --function-id='iii::queue::redrive' --payload='{"queue": "payment"}' --timeout-ms=60000 — with custom timeout
• Redrive returns {queue: 'payment', redriven: 12} indicating count of replayed jobs
• Inspect in RabbitMQ UI at http://localhost:15672, find iii.__fn_queue::{name}::dlq.queue
• Best practice: investigate failures, deploy fix, then redrive
• Monitor DLQ depth as an operational alert signal

Adapting This Pattern

Use the adaptations below when they apply to the task.

• Set max_retries and backoff_ms in queue_configs based on your failure tolerance
• Build an admin endpoint that calls iii::queue::redrive for operational control
• Use iii::queue::status to check DLQ depth before and after redriving
• For dev/test, use lower retry counts to surface failures faster
• In production with RabbitMQ, use the management UI for detailed message inspection
• Consider building an alerting function that triggers on DLQ depth thresholds

Engine Configuration

Queue max_retries and backoff_ms are set per queue in iii-config.yaml under queue_configs. See ../references/iii-config.yaml for the full annotated config reference.

Pattern Boundaries

• For queue processing patterns (enqueue, concurrency, FIFO), prefer iii-queue-processing.
• For queue configuration (retries, backoff, adapters), prefer iii-engine-config.
• For function registration and triggers, prefer iii-functions-and-triggers.
• Stay with iii-dead-letter-queues when the primary problem is inspecting or redriving failed jobs.

When to Use

• Use this skill when the task is primarily about iii-dead-letter-queues in the iii engine.
• Triggers when the request directly asks for this pattern or an equivalent implementation.

Boundaries

• Never use this skill as a generic fallback for unrelated tasks.
• You must not apply this skill when a more specific iii skill is a better fit.
• Always verify environment and safety constraints before applying examples from this skill.