CloudWatch vs CloudTrail: Differences and Which the AWS Exam Expects (2026)

Illustration of a monitor with a glowing graph beside a magnifying glass following a trail of footprints

Amazon CloudWatch watches how your resources and applications are performing: metrics, logs, alarms and dashboards. AWS CloudTrail records who did what in your AWS account: every API call, with the caller, time and source. On the exam, “monitor”, “metric”, “alarm” or “threshold” points to CloudWatch, and “audit”, “who made the change” or “API activity” points to CloudTrail. If the question asks what a resource’s configuration looked like over time, the answer is a third service, AWS Config.

Facts verified November 2026 against AWS documentation.

CloudWatch vs CloudTrail at a glance

In short: CloudWatch tells you how things are running, CloudTrail tells you who did what, and AWS Config tells you how resources were configured.

Amazon CloudWatchAWS CloudTrail
Main question it answers“How is it performing? Is something wrong?”“Who did what, when and from where?”
What it collectsMetrics (CPU, latency, request counts, custom metrics), logs from applications and services, eventsAPI calls and account activity: console sign-ins, CLI and SDK calls, actions by AWS services
Key featuresAlarms, dashboards, CloudWatch Logs and Logs Insights, the CloudWatch agentEvent history, trails delivered to S3, CloudTrail Lake, CloudTrail Insights, log file integrity validation
Default behaviourMany AWS services publish basic metrics automatically. EC2 sends metrics every 5 minutes by default, or every minute with detailed monitoringEvent history keeps 90 days of management events at no charge, without any setup
Long-term storageCloudWatch Logs with a retention setting you chooseA trail delivers log files to an S3 bucket you own, for as long as you keep them
Acts on what it seesYes: alarms can notify through SNS, trigger Auto Scaling or run EC2 and Systems Manager actionsNo direct actions. Its events can be routed to EventBridge or CloudWatch Logs, which then act
Typical userOperations and developers watching health and performanceSecurity, audit and compliance teams

Choose CloudWatch when…

  • You need a metric or a threshold: CPU above 80%, too many 5xx errors, a queue that’s backing up.
  • Something should happen automatically when a value crosses a line, such as notifying the team or adding instances.
  • You want application or system logs in one place, searchable with CloudWatch Logs Insights.
  • You need memory or disk-space metrics from EC2. Install the CloudWatch agent, because EC2 doesn’t send these by default.

Choose CloudTrail when…

  • You need to know who deleted, changed or created something, for example who terminated an instance or changed a security group.
  • An auditor wants a record of account activity that’s stored durably and can be checked for tampering. Log file integrity validation proves the files weren’t changed.
  • You manage many accounts and want one organisation trail that sends every account’s activity to a central bucket.
  • You want to spot unusual API activity, such as a sudden burst of resource creation. That’s CloudTrail Insights.

What about AWS Config?

AWS Config is the third service in this group: it records the configuration of your resources over time and checks it against rules. Config can show what a security group looked like last Tuesday and whether an S3 bucket is compliant with a “no public access” rule. CloudTrail shows the API call that changed it, and who made it. Questions often place all three in the options, so learn the split:

Question wordingService
“CPU utilisation”, “alarm when…”, “application logs”CloudWatch
“Who made this API call?”, “audit user activity”CloudTrail
“Configuration history”, “is this resource compliant?”, “detect configuration drift from a rule”AWS Config

Exam cues: which answer the question wants

Cue in the questionLikely answerExam
“Monitor resource utilisation”, “collect metrics”Amazon CloudWatchCLF-C02
“Track user activity and API usage”, “audit”AWS CloudTrailCLF-C02
“Notify the team when CPU exceeds a threshold”CloudWatch alarm → SNS topicCLF-C02, SOA-C03
“Find out who stopped an EC2 instance last week”CloudTrail event historyCLF-C02, SOA-C03
“Collect memory utilisation from EC2 instances”CloudWatch agentSOA-C03
“Keep API logs for 7 years for compliance”A CloudTrail trail delivering to S3 (with lifecycle rules)SOA-C03, SAA-C03
“Prove the log files haven’t been altered”CloudTrail log file integrity validationSOA-C03
“Capture activity from every account in the organisation”An organisation trail in AWS OrganizationsSOA-C03, SAA-C03
“Record object-level S3 reads and writes”CloudTrail data events (charged separately)SOA-C03
“Search and analyse application logs with queries”CloudWatch Logs InsightsSOA-C03
“Automatically fix a non-compliant resource”AWS Config rule with remediationSOA-C03

Worked example 1: the missing instance

An EC2 instance in production was terminated over the weekend. The operations manager needs to know which user terminated it and when. What is the quickest way to find out?

The question is about an action and the person who did it, not about performance. CloudWatch would show that the instance’s metrics stopped, but not who stopped it. AWS Config would show that the instance no longer exists, and it can link to the related API event, but it isn’t the direct record of the caller. CloudTrail records the TerminateInstances call, including the identity, time and source IP. Since it happened within the last 90 days, it’s already in CloudTrail event history, with no trail needed. The answer is to look up the event in the CloudTrail console’s event history.

Worked example 2: the slow web tier

A web application on an Auto Scaling group of EC2 instances becomes slow every afternoon. The operations team wants to be notified and to have capacity added automatically when average CPU stays above 70% for 10 minutes. What should they configure?

Every word here is about performance: “average CPU”, “above 70%”, “10 minutes”, “add capacity”. That’s CloudWatch. A CloudWatch alarm on the group’s average CPUUtilization metric, with a period and evaluation that cover 10 minutes, can do two things at once: publish to an SNS topic to notify the team, and trigger a scaling policy that adds instances. A target tracking policy is an even simpler way to keep CPU near a target. CloudTrail plays no part here, because nothing in the question is about who did something.

Common CloudWatch vs CloudTrail traps

  1. “CloudTrail monitors performance.” It doesn’t. It records API activity, so performance is always CloudWatch.
  2. “CloudWatch shows who made a change.” It doesn’t, unless you’ve sent CloudTrail events into CloudWatch Logs. The record of the caller comes from CloudTrail.
  3. Forgetting AWS Config. “What did this resource look like?” and “Is it compliant?” are Config questions.
  4. Memory metrics on EC2. They need the CloudWatch agent. Basic EC2 metrics don’t include memory.
  5. CloudTrail event history isn’t forever. It keeps 90 days of management events. For longer retention, data events or a central copy, create a trail to S3 or use CloudTrail Lake.
  6. CloudTrail isn’t real-time alerting on its own. To act on a specific API call, route the event to EventBridge or CloudWatch Logs and build the alert there.

CloudWatch vs CloudTrail FAQ

Test yourself

Monitoring and auditing questions appear on every AWS exam. Try them in our free AWS Cloud Practitioner practice exam, then take a full timed test.

Free mock exam

Take CLF-C02 Mock Exam 1, free

A full-length, timed mock in the real exam format, with every answer explained so you know exactly where you stand.

65

questions

90

minutes

700

to pass

Free account required.

Studying for CloudOps? Our SOA-C03 practice exams go deeper, and SOA Mock Exam 1 is free too (free account required).

Read the CLF-C02 exam guide or the SOA-C03 exam guide.


AWSBoy is not affiliated with Amazon Web Services. AWS and related marks are trademarks of Amazon.com, Inc. or its affiliates.

Sources (checked November 2026): Amazon CloudWatch User Guide · Installing the CloudWatch agent · AWS CloudTrail User Guide · Working with CloudTrail event history · AWS CloudTrail pricing · AWS Config Developer Guide

Shopping Cart
Scroll to Top