Priority‑based message processing system that routes urgent tasks instantly while handling routine work reliably.

Infrastructure as Code with AWS CDK

Using the AWS Cloud Development Kit lets you describe every resource in code, making environments reproducible and version‑controlled. The stack provisions encryption, monitoring, and networking in a single command.

• Python‑based CDK app initializes with cdk init app --language python.
• Creates a KMS key for data‑at‑rest encryption.
• Defines a DynamoDB table with on‑demand billing and streams enabled.
• Deploys an Amazon MQ broker configured for ActiveMQ engine.
• Outputs critical identifiers (table name, broker endpoint) for application wiring.

Priority Queue Configuration in Amazon MQ

Amazon MQ supports JMS priority levels, allowing the broker to order messages without custom code. High‑priority messages (9) bypass delay logic, while lower levels wait their turn.

• Set JMS priority values High = 9, Medium = 4, Low = 0 (see JMS spec).
• Configure dead‑letter queue (DLQ) for messages that exceed retry limits.
• Enable persistent delivery mode to guarantee storage until acknowledged.
• Use broker‑level access control with dedicated user credentials.
• Monitor queue depth and latency via Amazon CloudWatch metrics.

Delay Management and Asynchronous Processing

Routine messages enter a non‑blocking delay pipeline before reaching the broker, freeing compute resources for urgent work. The implementation relies on Java CompletableFuture and a managed thread pool.

• Define per‑priority delay intervals (e.g., 30 s for Medium, 2 min for Low).
• Execute delays with CompletableFuture.runAsync to avoid thread starvation.
• Adjust thread‑pool size based on expected throughput.
• Log start‑ and end‑times for auditability.
• Integrate accelerated JavaScript runtime for lightweight side‑car services.

Real‑time Status Updates via WebSocket and DynamoDB Streams

Clients receive live processing feedback through a WebSocket API powered by AWS App Runner. DynamoDB Streams capture every table change and push updates downstream.

• App Runner hosts a lightweight Spring/React gateway exposing /ws endpoint.
• Subscribe to DynamoDB Streams on INSERT or MODIFY events, broadcast status via WebSocket.
• Frontend React component consumes messages and renders a progress bar.
• Implement back‑pressure handling to protect the WebSocket server.
• Reference Google I/O 2026 insights for UI patterns.

Retry Logic and Dead‑Letter Handling

Automatic retries improve reliability, while a dead‑letter queue isolates permanently failing messages for later analysis. Both layers are configurable per priority level.

• First‑level retry uses exponential back‑off (e.g., 1 s, 2 s, 4 s).
• Second‑level retry moves the message to the DLQ after N attempts.
• DLQ is a separate Amazon MQ queue monitored by an alerting Lambda.
• Store failure reason and metadata in DynamoDB for forensic review.
• Expose metrics to CloudWatch for SLA reporting.