Event-Driven GenAI Workflow with S3, EventBridge, and Lambda
Short answer
CD Use EventBridge to trigger Step Functions (D) and Lambda to handle prompt construction and Bedrock invocation (C), bypassing S3 direct integration limits and Step Functions payload constraints.
Question
Which combination of solutions will meet these requirements? (Choose two.)
Key Takeaways
The main trap is assuming S3 can trigger Step Functions directly or that Step Functions can handle large transcripts (>256 KiB) without offloading logic to Lambda.
This question tests designing an event-driven architecture for processing large audio files with Amazon Bedrock. The correct solution uses EventBridge to trigger Step Functions and Lambda to manage payload limits during FM invocation.
Common mistake: Choosing B and D or A and D. Users often overlook the 256 KiB payload limit in Step Functions, leading them to believe Bedrock can be invoked directly with large transcript text, or they incorrectly assume S3 supports direct Step Functions notifications.
Expert Analysis
Triggering Mechanism: The requirement is to react "as soon as new files are created" in an S3 bucket. AWS S3 does not send notifications directly to Step Functions. Instead, the standard pattern is to configure S3 Event Notifications to Amazon EventBridge (or SQS/SNS), which then triggers the Step Functions state machine. Option D correctly describes this event-driven architecture using EventBridge. Option E is technically invalid because S3 cannot invoke Step Functions directly without a bridge service like EventBridge or SNS.
Workflow Logic and Payload Limits: Once triggered, the workflow must transcribe audio and generate structured output. A critical constraint here is the size of the recordings (20 MB MP3s). While Amazon Transcribe outputs text, large transcripts can easily exceed the 256 KiB payload limit imposed on AWS Step Functions state data. Passing the full transcript directly between states or invoking Bedrock from within a Step Function state using the raw transcript text risks hitting these limits. Option C addresses this by using an AWS Lambda function. Lambda has significantly higher memory and execution time limits, allowing it to safely retrieve the transcript from S3, construct a complex prompt for sentiment analysis, and invoke Bedrock's API to return structured JSON. Option B suggests invoking Bedrock directly from Step Functions; while possible for small inputs, it is risky for large transcripts and less flexible for prompt engineering than Lambda. Option A adds unnecessary complexity by having Lambda process the output after Bedrock, whereas the goal is generation.
Conclusion: Options C and D form a robust, scalable solution that handles the event trigger correctly and manages data payload constraints via Lambda.
Official Reference
- https://docs.aws.amazon.com/step-functions/latest/dg/connect-s3.html
- https://docs.aws.amazon.com/AmazonS3/latest/userguide/ev-events.html
- https://docs.aws.amazon.com/bedrock/latest/userguide/model-limits.html
These links open the official vendor documentation in a new tab.
Exam Strategy
Always check for "direct" integrations between services that require a bridge (like S3 -> Step Functions). Additionally, remember Step Functions' 256 KiB payload limit; if data might be large, use Lambda to fetch/process data from S3 rather than passing it through state variables.
Frequently Asked Questions
Can S3 send notifications directly to AWS Step Functions?
Why use Lambda instead of calling Bedrock directly from Step Functions?
Community Votes
50% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.
Discussion 7 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 7
-
ericcloud20 👍 1Why B: the processing Transcribe step: Bedrock models analyze text, not MP3 audio, so the recording has to be transcribed first. Step Functions starts the Transcribe job and loops to check its status until the job is done. Calling Bedrock directly: Step Functions has a built-in integration for Bedrock InvokeModel. The state machine can build the prompt from the transcript itself, asking for a JSON object with summary and sentiment fields, and call the model with no Lambda in between. That’s less code to write, deploy, and maintain. Size limits: the 20 MB MP3 goes to Transcribe by its S3 location, not through the workflow payload, so its size doesn’t matter. The Bedrock integration can also read model input from an S3 location, so a long transcript doesn’t run into Step Functions’ state payload size limit. Why D: the trigger S3 event notifications can only go to four destinations: SNS topics, SQS queues, Lambda, and EventBridge. Step Functions isn’t one of them. Sending events to EventBridge with a rule that targets the state machine is the standard way to start a workflow “as soon as new files are created.”
-
Chibuzo1 👍 1The answers are C and D because together they form a complete, production-ready solution: EventBridge detects new MP3 files and triggers Step Functions (D), which orchestrates Transcribe for audio-to-text conversion, uses Lambda to construct a structured prompt from the transcript, and invokes Bedrock FMs to generate consistent JSON-formatted summaries and sentiment analysis (C) — all fully event-driven with proper async job handling and no manual intervention.
-
Wardove 👍 1CD wins because Transcribe outputs JSON with metadata (not a ready Bedrock prompt) and transcripts from 20 MB MP3s exceed Step Functions' 256 KiB state payload limit — so Lambda is required to read the transcript from S3, craft a proper prompt, and invoke Bedrock internally (C), while S3 must route events through EventBridge since Step Functions isn't a direct S3 notification destination (D).
-
2d5eb96 👍 1Option A provides complete orchestration with Lambda for prompt engineering, Bedrock invocation, and structured output processing Option D uses EventBridge for reliable, scalable event routing from S3 to Step Functions with advanced filtering capabilities
-
Oldgun99 👍 1Amazon Transcribe produces raw transcript text Raw transcript text cannot be directly passed to Bedrock FMs and reliably produce structured output — it requires a carefully engineered prompt
Show 2 more comments
-
nebulane 👍 1CD because you need to prompt the FM to generate the structured data aside from the raw mp3 text
-
de1612d 👍 1Gemini answer is A,D
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.
Related Analysis
Practice All AIP-C01 Questions
Access 110 questions with complete answers and detailed explanations.
View Full AIP-C01 Practice Test →