> ## Documentation Index
> Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Demo Creating and Configuring Alarms with alerting
> This hands-on tutorial teaches launching an EC2 instance, generating load, and setting up CloudWatch alarms with SNS notifications.
In this hands-on tutorial, you’ll learn how to launch an EC2 instance, generate system load, and set up a CloudWatch alarm with SNS notifications. By the end, you’ll have:
* A running EC2 instance under stress
* A CPU utilization alarm
* Email notifications on alarm trigger
## 1. Launch an EC2 Instance
1. Sign in to the AWS Management Console and confirm you’re in the correct region.
2. Navigate to **EC2 → Instances** and click **Launch Instance**.
3. Under **Name**, enter `demo-instance`. Select **Amazon Linux 2023** as your AMI.
4. In **Key pairs**, choose **Proceed without a key pair**.
Launching without a key pair prevents SSH login via key-based authentication. You’ll connect using EC2 Instance Connect or Session Manager.
5. Keep the default VPC and subnet. Under **Security groups**, select **Create security group** with default rules.
6. Click **Launch instances**, then view your instance on the Instances page. Wait until its status changes from **Pending** to **Running**.
## 2. Connect to the Instance
1. Select your running instance and click **Connect**.
2. Choose **EC2 Instance Connect** and then **Connect** again to open the browser-based shell.
3. Switch to root:
```bash theme={null}
[ec2-user@ip-172-31-16-22 ~]$ sudo su -
[root@ip-172-31-16-22 ~]#
```
## 3. Install and Run the stress Utility
CloudWatch doesn’t collect detailed CPU or memory load by default. To simulate load:
```bash theme={null}
[root@ip-172-31-16-22 ~]# yum install stress -y
```
Once installed, launch `stress`:
```bash theme={null}
[root@ip-172-31-16-22 ~]# stress --cpu 4 --io 4 --vm-bytes 1G --hdd 2
stress: info: [25930] dispatching hogs: 4 cpu, 4 io, 1 vm, 2 hdd
```
Use `Ctrl+C` to stop the load.
### stress Command Options
| Flag | Description | Example |
| ---------- | --------------------- | --------------- |
| --cpu | Number of CPU workers | `--cpu 4` |
| --io | Number of I/O workers | `--io 4` |
| --vm-bytes | Memory per VM worker | `--vm-bytes 1G` |
| --hdd | Number of HDD workers | `--hdd 2` |
## 4. View Metrics in CloudWatch
1. Open the **CloudWatch** console in a new tab.
2. Go to **Alarms → Create an alarm**.
3. Click **Select metric**, navigate to **EC2 → Per-Instance Metrics**, and search your instance ID.
4. Initially select **CPU Credit Balance**, **CPU Credit Usage**, and **CPU Utilization**, then set the graph to a 1-hour view to spot spikes.
5. Because alarms support only one metric at a time, uncheck the credit metrics and keep **CPU Utilization** selected, then click **Select metric**.
## 5. Configure Alarm Conditions
1. Under **Threshold type**, choose **Static**.
2. Set **Greater than or equal to 65** (%).
3. Click **Next**.
You can also use **Anomaly detection** for adaptive thresholds that adjust based on historical data.
## 6. Set Alarm Actions
1. Under **Select an SNS topic**, choose an existing topic—e.g., `CloudWatch-course-topic`.
2. This topic delivers email alerts to its subscribers.
3. Click **Next**.
4. Provide an alarm name (e.g., `cpu-utilization-alarm`) and an optional message. Preview the email content.
5. Click **Next**, then **Create alarm**. Initially, the alarm shows **Insufficient data**; once metrics arrive and CPU ≥ 65%, it shifts to **In alarm**.
## 7. Verify Notifications
Check your email inbox for an alert:
Email contents include:
* Alarm name
* Instance details
* Custom message
* CloudWatch console link
* Triggering metric graph
## 8. Resolve the Alarm
1. On the EC2 instance, stop `stress` with `Ctrl+C`.
2. After CPU drops below 65%, the alarm transitions to **OK**.
Note that CloudWatch doesn’t notify on resolution by default. To get recovery alerts, edit the alarm’s actions or create a second alarm for the OK state.
## Summary
You’ve successfully:
* Launched and connected to an EC2 instance
* Installed and ran `stress` to generate CPU load
* Navigated CloudWatch and created a CPU utilization alarm
* Configured SNS notifications and verified delivery
* Resolved and monitored alarm recovery
## References
* [Amazon EC2 Documentation](https://docs.aws.amazon.com/ec2/)
* [Amazon CloudWatch User Guide](https://docs.aws.amazon.com/cloudwatch/)
* [stress on GitHub](https://github.com/stress-ng/stress-ng)