What is a dead-letter queue and when would you use it in CDX?

• A. A queue for messages that fail processing; used to isolate, analyze, and retry problematic data
• B. A queue for prioritized messages
• C. A backup storage for all messages
• D. A mechanism to permanently delete messages after processing

Answer: (A)

A dead-letter queue is a separate holding area for messages that can’t be processed after repeated attempts, allowing isolation and troubleshooting without blocking the main processing flow. In CDX, when a consumer encounters an error—such as malformed data, missing fields, or a schema mismatch—and retries fail beyond a configured limit, that problematic message is moved to the dead-letter queue. This preserves the good messages in the primary queue while giving operators a place to analyze, diagnose the root cause, and fix or reprocess the data once the issue is resolved. It also supports alerting and targeted remediation workflows so the system remains healthy and processing throughput isn’t disrupted.

This isn’t a general backup for all messages, nor a place for prioritizing data, and it isn’t meant for permanently deleting messages after processing. Rather, it serves as a dedicated space for the anomalies that block normal processing, so they can be addressed without impacting the rest of the pipeline.