This article explains how to stream data changes from Amazon DocumentDB to Amazon S3 using AWS Lambda and Kinesis Data Firehose (now Amazon Data Firehose).

• Lambda reads DocumentDB change stream events via event source mappings
• Lambda function processes changes and pushes data to Firehose delivery stream
• Firehose automatically archives data to S3 with automatic prefixing by date/hour
• Setup requires DocumentDB cluster, Lambda function, S3 bucket, and Firehose delivery stream
• Change streams must be enabled on DocumentDB collections to capture insert/update/delete operations
• Lambda needs VPC endpoints for DocumentDB and Secrets Manager connectivity
• Secrets Manager stores DocumentDB credentials for secure Lambda access
• UpdateLookup option captures full document state but has performance overhead
• Firehose has limits: 2,000 transactions/sec, 5,000 records/sec, 5 MB/sec per stream
• Solution enables archiving, backup, and analytics use cases with minimal code

This approach provides a serverless, low-code solution for continuously archiving DocumentDB changes to S3 for analytics, backup, or data warehouse integration.