AWS CloudWatch
AWS's observability service — metrics, logs, alarms, dashboards, traces (via X-Ray), RUM, synthetic monitoring. Billed per metric, log GB and dashboard.
Logs Insights (interactive query language over logs) is its bright spot. Metrics resolution: 1 min default, 1 sec for "high-resolution". Composite alarms (combine multiple alarms with boolean logic) since 2020. Container Insights for ECS/EKS.
Criticised: expensive at high-cardinality metrics, clunky UX. Most serious shops use Datadog/Grafana/Honeycomb on top for real incident investigations. CloudWatch lives for billing-default alerts and AWS services' own metrics.