CloudWatch alarm uses Average statistic for a latency-shaped metric
What it checks
CloudWatch alarm uses Average statistic for a latency-shaped metric
Where it applies
Terraform: aws_cloudwatch_metric_alarm
CloudFormation: AWS::CloudWatch::Alarm
How to fix it
- Confirm existing mean and percentile objectives. If tail coverage is needed, add a separate p95/p99 alarm with a threshold derived from the workload SLO.
Frameworks
- AWS Well-Architected OPS.8
These mappings are advisory, not an attestation of compliance.