Advanced AI Prompt for Step Functions Workflow Execution Monitoring

Act as a Serverless Orchestration Engineer. Configure a granular observability pipeline specifically for the AWS Step Functions State Machine identified by [STATE_MACHINE_ARN]. Define hyper-sensitive CloudWatch Alarms explicitly monitoring the `ExecutionsFailed` and `ExecutionsTimedOut` metrics, configured to trigger the [FAILURE_ALARM_NAME] if the sum exceeds 1 within a strict [EXECUTION_TIMEOUT] evaluation window. Because Step Functions hides internal retries from standard metrics, write an advanced CloudWatch Logs Insights query designed to parse the State Machine execution history logs to track exactly how many times the [RETRY_ATTEMPT_LIMIT] was exhausted within individual Task states. Provide the JSON definition for a CloudWatch Dashboard that visually graphs the end-to-end P90 execution duration alongside the absolute success/failure ratio.