# CloudWatch Alarms at a Glance — AWS

Source: https://www.geekswithgeeks.com/en/aws/aws-cloudwatch-alarms

> Turning a bad metric into an automatic notification or action.

## What an alarm is

A **CloudWatch alarm** watches one metric and triggers when it crosses a threshold — e.g. CPU above 80% for 5 minutes — moving you from checking dashboards to being notified.

## Alarms can act, not just alert

An alarm can send a notification (via SNS to email/Slack) or trigger an action directly, like scaling out an Auto Scaling Group when load is high.

## Set alarms before you need them

Add alarms for CPU, error rates, and queue depth on day one — finding out about an outage from a customer instead of an alarm is a bad first sign.

**Quiz:** What can a CloudWatch alarm do besides sending a notification?

- [ ] Nothing, it only notifies
- [x] Trigger an action, like Auto Scaling
- [ ] Automatically fix application bugs

*Answer:* Trigger an action, like Auto Scaling. Alarms can trigger Auto Scaling actions or other automated responses, not just notify.