Question

A company runs a customer-facing web application. The operations team wants to be alerted within minutes whenever the login workflow becomes unavailable, even if no real users are attempting to log in at that moment (for example, during off-peak overnight hours). They need to measure the availability and latency of a specific multi-step user journey from an external perspective, and trigger an SNS notification when the journey fails. Which approach best meets these requirements with the least custom development?

Accepted Answer

Create a CloudWatch Synthetics canary that runs a script simulating the login workflow on a schedule, then create a CloudWatch alarm on the canary's SuccessPercent metric that notifies an SNS topic.

Answer

Enable AWS X-Ray on the application and create a CloudWatch alarm on the X-Ray error trace count, sending notifications to an SNS topic.

Answer

Create a CloudWatch Logs metric filter on the application's access logs that counts HTTP 200 responses to the login endpoint, and alarm when the count drops to zero.

Answer

Configure an EventBridge rule that matches ALB target-group health-check failure events and routes them to an SNS topic.