We’re pushing routing.interaction.created events to a Lambda consumer. Volume spikes around 500 TPS during peak. Lambda keeps hitting the concurrency limit and dropping events.
Currently using:
{"target": "arn:aws:lambda:us-east-1:123:func:consumer", "batchSize": 100}
Increasing batchSize helps slightly but not enough. Any code pattern to handle backpressure or fan out to multiple handlers without manual queue management? Just need a clean way to absorb the burst.