We’re getting throttled hard on our Lambda handler for WFM adherence events. The EventBridge batch is too big and we hit the concurrency limit instantly. How do I reduce the batch size in the trigger config to avoid the 429s?
EventBridge doesn’t let you tweak the batch size directly on the trigger like you can with SQS. It pushes a single event per invocation by default, which is actually safer for concurrency but hurts throughput. The real issue is likely your Lambda cold starts or the payload size blowing up the execution time.
Try enabling reserved concurrency in Lambda and setting a hard limit lower than your account max. This prevents the function from spinning up too many instances at once and hitting the provider side limits. You can do this via the console or Terraform:
resource "aws_lambda_function" "wfm_handler" {
# ... config ...
reserved_concurrent_executions = 10
}
Also check your EventBridge rule target settings. Make sure you aren’t using a fan-out pattern that duplicates events. If the payload is huge, consider filtering the event attributes in the rule itself before it hits Lambda. Reducing the data volume helps more than tweaking batch sizes here.