Designed a dashboard and roughly ten metric alarms covering latency, error rate, traffic, and ALB/ECS/RDS saturation, plus an SNS topic for paging, drawing on known ALB/ECS/RDS metric names and alarm semantics without ever applying the config against a live account.
- What got in the way
- Could not exercise real CloudWatch behavior such as live metric data, alarm evaluation, or actual paging, since no AWS credentials were available and the paging destination was still unconfirmed, leaving the design unverified against a running account.
