Observability: metrics, logs, traces · Lesson 5 of 7 · 3 min read
AWS-native observability
- CloudWatch Metrics: namespaces, dimensions, standard (1-minute) and high-resolution (1-second) custom metrics, metric math, Embedded Metric Format (emit metrics through structured logs from Lambda/containers).
- CloudWatch Alarms: static thresholds, anomaly detection bands, composite alarms (reduce noise: page only if latency AND errors),
TreatMissingDatachoices, actions to SNS/Auto Scaling/EC2. - Container Insights (EKS/ECS) with enhanced observability: cluster/node/pod/container metrics and dashboards via the CloudWatch agent add-on.
- CloudWatch Logs: retention per log group, metric filters, subscription filters to Lambda/Firehose/OpenSearch, Logs Insights, Live Tail.
- X-Ray / Application Signals: tracing and service-level dashboards with SLOs.
- Synthetics canaries (scripted user journeys) and RUM (real-user monitoring).
- Managed Prometheus (AMP) + Managed Grafana (AMG) for Prometheus-native teams.
- Load balancer metrics (
HTTPCode_Target_5XX_Count,TargetResponseTime,UnHealthyHostCount), access logs to S3 — first stop for edge issues.
Enlarge diagram
Swipe sideways to see all of it, or tap to open it full screen.