Question

A GenAI customer-support application invokes an Amazon Bedrock model through a Lambda function. Traffic follows a strong daily pattern: high volume during business hours and near-zero overnight. The team wants to be alerted when the application's error rate (failed invocations) deviates significantly from its normal behavior, but a static CloudWatch alarm threshold keeps producing false alarms overnight (when a single error looks like a 100% error rate) and misses moderate spikes during peak load. What is the MOST appropriate way to configure the alerting?

Accepted Answer

Create a CloudWatch anomaly detection alarm on the error-rate metric so the expected band adapts to the daily traffic pattern, and alarm when values fall outside the band.

Answer

Lower the static threshold so that even a single overnight error stays below it, accepting that peak-hour spikes will trigger alarms.

Answer

Configure a metric math expression that only evaluates the alarm during business hours and disables it overnight with a scheduled Lambda.

Answer

Switch the alarm to evaluate the raw error count with a high threshold so overnight single errors never trigger it.