Deployment Architecture (AWS SAM)

This kit deploys with AWS SAM into a single stack:

Resources

API Gateway

REST API with two endpoints:

• POST /ingest → Ingest Lambda
• POST /replay → Replay Lambda

Storage

• S3 Bucket: Immutable raw event payloads
• DynamoDB Tables:
  • EventsTable: Metadata and index records
  • DedupeTable: Idempotency gate for duplicate detection
  • AggregatesTable: Versioned aggregate results (optional, only when EnableProcessor=true)

Messaging

• SQS Queue: ReplayQueue for event reprocessing

Compute

Lambda Functions:

• IngestFunction: Validates and stores incoming events
• ReplayFunction: Queries and replays historical events
• ProcessorFunction: Consumes and processes events from SQS (optional)

Parameters

• EnableProcessor (bool)
• false: no processor / no aggregates table / no event source mapping
• true: creates Processor + AggregatesTable and wires SQS → Processor

Environments

Each Lambda receives env vars for resource names/urls:

• RAW_BUCKET
• EVENTS_TABLE
• DEDUPE_TABLE
• REPLAY_QUEUE_URL
• AGG_TABLE (processor only)
• LOG_LEVEL

IAM

Each Lambda has a dedicated role with least-privilege access:

• Ingest: PutObject to raw bucket, PutItem to tables, PutMetricData/logging
• Replay: Query EventsTable, SendMessage to SQS
• Processor: GetObject from raw bucket, PutItem to AggregatesTable

Components in this stack

• API Gateway (REST)
• Lambda (Ingest, Replay, optional Processor)
• DynamoDB (EventsTable, DedupeTable, optional AggregatesTable)
• S3 (raw immutable storage)
• SQS (ReplayQueue)
• CloudWatch (logs + EMF metrics)
• IAM (least privilege roles)
• X-Ray (tracing)