I have a standard CW alarm --> SNS --> Lambda
set up. The alarm console shows that it executed 1 successful action to the SNS topic. But on the lambda console it triggered 3 times, with the same request ID and payload? Only the execution times are different. Any help appreciated :)
Your description matches the behavior of retry submissions from SNS to Lambda. Basically, SNS will try to retry 3 times without any delay on a failed submission.
Thus it seems that for some reason your lambda is failing and SNS immediately retries. Please inspect cloudwatch logs for possible error messages.