When your app in production has errors or gets slow, you need to find out what went wrong. This capability is called Observability. The AWS DVA-C02 exam asks which tools to use to identify the root cause.
Amazon CloudWatch
CloudWatch is AWS's central monitoring service — a control tower watching all your services in one place. It has five main capabilities.
CloudWatch Logs
Collects logs from all AWS services in one place. Log Group: A logical container for organizing similar logs, like a folder. Log Stream: A time-ordered sequence of logs from a specific source. Retention Period: From 1 day to permanent (default: permanent).
CloudWatch Metrics
Numerical measurements of system health: CPU usage, network traffic, request counts. Collected typically every 1 minute.
Think of it like a car's speedometer, fuel gauge, and engine temperature gauge.
CloudWatch Logs Insights
An analysis tool for quickly finding specific information in large log volumes using SQL-like syntax. For example: "Show me only error logs from the past hour."
CloudWatch Alarms
Automatically sends notifications or triggers actions when a metric crosses a threshold you set. For example, if CPU usage exceeds 80%, send an SMS or automatically add servers.
CloudWatch Dashboards
Visualizes multiple metrics and logs as graphs on a single screen so the whole team can see service health at a glance.
AWS X-Ray
X-Ray traces the complete path of requests as they travel through multiple distributed services. Think of it like a package tracking system showing exactly how long a delivery spent at each distribution center.
Service Map: Shows relationships between services and the latency of each segment visually. Trace: The full recorded path of one request through multiple services. Segment: The time a specific service took to handle its part of the request. Subsegment: Detailed operations within a segment (e.g., DB query time, external API call time).
X-Ray lets you immediately see which service is the bottleneck slowing everything down.
EMF (Embedded Metric Format)
With EMF, CloudWatch automatically extracts metrics from logs when they are output in a special JSON format — no separate PutMetricData API call needed.
Exam Key Points
"Analyze logs with SQL" -- CloudWatch Logs Insights
"Trace request flow in distributed apps" -- X-Ray
"Visualize service connections and latency" -- X-Ray Service Map
"Alert when metric crosses threshold" -- CloudWatch Alarms
"Auto-extract metrics from structured logs" -- EMF (Embedded Metric Format)
"View multiple metrics on one screen" -- CloudWatch Dashboards
"Identify bottlenecks in request path" -- X-Ray traces and segments
X-Ray integrates with Lambda, API Gateway, ECS, and more