RDS Alarm Verification¶
Date: 2026-06-25 Status: Accepted for the first paid pilot. Roadmap item: EP8-T5.
Decision¶
Production RDS alarm coverage is accepted for the first paid pilot.
The architecture runbook now gives Product Owner and Infra Developer a direct database health lookup path:
aws cloudwatch describe-alarms --alarm-name-prefix mamnajem-production-rds --region eu-west-1
Verified Coverage¶
| Health area | Production alarm |
|---|---|
| CPU | mamnajem-production-rds-cpu-high |
| Storage | mamnajem-production-rds-free-storage-low |
| Memory | mamnajem-production-rds-freeable-memory-low |
| Connections | mamnajem-production-rds-connections-high |
| Read latency | mamnajem-production-rds-read-latency-high |
| Write latency | mamnajem-production-rds-write-latency-high |
| Disk queue depth | mamnajem-production-rds-disk-queue-depth-high |
| Burst balance | mamnajem-production-rds-burst-balance-low |
CPU, storage, memory, connections, and latency meet the EP8-T5 launch requirement. Disk queue depth and burst balance are additional early-warning signals for storage pressure.
Evidence¶
- Architecture Terraform:
terraform/aws-production/alerting.tf. - Architecture runbook:
docs/runbooks/aws-alerts.md. - Architecture guard:
AWS observability Terraform > documents new alarm meanings, integration traffic checks, and remaining risks.
Remaining Boundary¶
This closes the RDS alarm discoverability gate for EP8-T5. Thresholds are initial conservative values and should be tuned after one week of production baseline data.