Observability for AWSs
AWS observability is seeing what your systems are doing and defining them repeatably. CloudWatch collects metrics, logs and alarms, CloudTrail records API activity for audit, X-Ray traces requests, and infrastructure as code tools such as CloudFormation and Terraform describe resources in files that can be reviewed and reused.
Stage: Operate · Part of the AWS roadmap · Last reviewed 19 November 2026
What to learn in Observability
- Create a CloudWatch alarm on a metric and send a notification
- Query application logs with CloudWatch Logs Insights
- Explain what CloudTrail records and why it matters
- Describe a stack of resources in a CloudFormation template
- Preview and apply infrastructure changes from code
Observability interview questions
What is the difference between CloudWatch and CloudTrail?
CloudWatch collects operational data such as metrics, logs and alarms to show how systems perform. CloudTrail records API calls made in your account, answering who did what, when and from where. You use CloudWatch for monitoring health and CloudTrail for auditing and security investigation.
Why use infrastructure as code instead of the console?
Code makes infrastructure repeatable, reviewable and version controlled. You can recreate environments identically, see changes in a diff before applying them and roll back. Console clicks are hard to reproduce and leave no record of exactly what was configured.
Where Observability fits in the AWS roadmap
Observability belongs to the Operate stage (advanced level), which takes about 25 hours. Monitor, secure and optimise.