[{"data":1,"prerenderedAt":81},["ShallowReactive",2],{"pattern-agentcore-memory-kinesis-typescript-cdk":3,"pattern-group-agentcore-memory-kinesis-typescript-cdk":6},{"id":4,"title":5,"architectureURL":6,"cleanup":7,"contributors":10,"deploy":12,"description":17,"extension":18,"framework":19,"gitHub":20,"highlight":6,"introBox":24,"language":31,"level":32,"meta":33,"patternArch":34,"patternGroup":6,"patternGroupLabel":6,"resources":62,"stem":76,"testing":77,"videoId":6,"__hash__":80},"patterns\u002Fpatterns\u002Fagentcore-memory-kinesis-typescript-cdk.json","Amazon Bedrock AgentCore Memory record streaming to Amazon Kinesis Data Streams",null,{"text":8},[9],"\u003Ccode>npx cdk destroy\u003C\u002Fcode>",[11],"content\u002Fcontributors\u002Fnithin-chandran-r.json",{"text":13},[14,15,16],"\u003Ccode>npm install\u003C\u002Fcode>","\u003Ccode>npx cdk bootstrap\u003C\u002Fcode> (once per account and Region)","\u003Ccode>npx cdk deploy\u003C\u002Fcode>","Stream Amazon Bedrock AgentCore Memory long-term records to Amazon Kinesis Data Streams and sync an Amazon DynamoDB table with AWS Lambda, including deletions.","json","AWS CDK",{"template":21},{"patternPath":22,"templateFile":23},"agentcore-memory-kinesis\u002Ftypescript\u002Fcdk","lib\u002Fagentcore-memory-kinesis-stack.ts",{"headline":25,"text":26},"How it works",[27,28,29,30],"Amazon Bedrock AgentCore Memory extracts long-term records with the built-in semantic and user-preference strategies and publishes a created, updated or deleted event for each record to Amazon Kinesis Data Streams.","An AWS Lambda event source mapping filters on the documented event types and skips the `StreamingEnabled` validation event.","The AWS Lambda function upserts or deletes an Amazon DynamoDB item keyed by memory record ID, so deletions reach the table even though deleted events carry no actor or namespace. If the function still fails on a record after its retries, the event source mapping sends metadata about the failed batch to an Amazon Simple Queue Service (Amazon SQS) queue.","The stream uses the `FULL_CONTENT` level, so the extracted memory text, which can include personal information from conversations, is copied into the stream and the Amazon DynamoDB table. Use `METADATA_ONLY` if you don't need the text.","TypeScript","300",{},{"icon1":35,"icon2":40,"icon3":44,"icon4":48,"line1":52,"line2":56,"line3":59},{"x":36,"y":37,"service":38,"label":39},12,50,"bedrock-agentcore","Amazon Bedrock\nAgentCore Memory",{"x":41,"y":37,"service":42,"label":43},38,"kinesis","Amazon Kinesis\nData Streams",{"x":45,"y":37,"service":46,"label":47},63,"lambda-event-functions","AWS Lambda\nRecord sync",{"x":49,"y":37,"service":50,"label":51},88,"dynamodb","Amazon DynamoDB\nMemory records",{"from":53,"to":54,"label":55},"icon1","icon2","Events",{"from":54,"to":57,"label":58},"icon3","Filtered",{"from":57,"to":60,"label":61},"icon4","Put, Delete",{"bullets":63},[64,67,70,73],{"text":65,"link":66},"Amazon Bedrock AgentCore Memory record streaming","https:\u002F\u002Fdocs.aws.amazon.com\u002Fbedrock-agentcore\u002Flatest\u002Fdevguide\u002Fmemory-record-streaming.html",{"text":68,"link":69},"Amazon Bedrock AgentCore Memory strategies","https:\u002F\u002Fdocs.aws.amazon.com\u002Fbedrock-agentcore\u002Flatest\u002Fdevguide\u002Fmemory-strategies.html",{"text":71,"link":72},"Using AWS Lambda with Amazon Kinesis Data Streams","https:\u002F\u002Fdocs.aws.amazon.com\u002Flambda\u002Flatest\u002Fdg\u002Fwith-kinesis.html",{"text":74,"link":75},"Retain discarded batch records for a Kinesis Data Streams event source in Lambda","https:\u002F\u002Fdocs.aws.amazon.com\u002Flambda\u002Flatest\u002Fdg\u002Fkinesis-on-failure-destination.html","patterns\u002Fagentcore-memory-kinesis-typescript-cdk",{"text":78},[79],"See the GitHub repo for detailed testing instructions.","lrsZOuVv6PLgWAod6zMkR8JeSMaQmiBg3nc2JH-EFlQ",1795084457782]