EC2 Monitoring and Alerts with AWS CloudWatch Agent, Alarms & Dashboards

Learn how to monitor an EC2 instance in real-time using Amazon CloudWatch Agent, create alarms for high CPU usage, and visualize everything with a custom dashboard.

---

🧭 Introduction

For my fourth DevOps project, I focused on observability — a crucial aspect of production readiness. This project demonstrates how to monitor an EC2 instance using CloudWatch Agent, create alarms, and build dashboards in AWS.

The result: a powerful, real-time monitoring setup with no third-party tools.

🔗 GitHub Repo: EC2 CloudWatch Monitoring Project

---

🎯 Objectives

• Monitor EC2 metrics: CPU, memory, disk usage

• Use CloudWatch Agent to push system-level data

• Create a CloudWatch Alarm for high CPU usage

• Visualize with a CloudWatch Dashboard

• (Optional) Send alarm notifications using SNS

---

🧱 Architecture

[EC2 Instance] ---> [CloudWatch Agent] ---> [CloudWatch Metrics]
                                       └---> [Logs / Alarms / Dashboards]
                                                    |
                                                    v
                                             [Optional: SNS Alert]

---

⚙️ Setup and Configuration

Step 1: Launch EC2 Instance

• Use Amazon Linux 2 AMI

• Assign IAM Role: CloudWatchAgentServerPolicy

Step 2: Install & Configure CloudWatch Agent

# SSH into the instance
scp cloudwatch-agent-config.json ec2-user@<EC2_PUBLIC_IP>:
scp cloudwatch-setup.sh ec2-user@<EC2_PUBLIC_IP>:
ssh ec2-user@<EC2_PUBLIC_IP>
chmod +x cloudwatch-setup.sh
sudo ./cloudwatch-setup.sh

📸 Screenshot: Agent installed and running

---

Step 3: Verify Metrics in CloudWatch

• Go to CloudWatch → Metrics → CWAgent → InstanceId

• Check graphs for:

  • cpu_usage_idle, cpu_usage_user

  • mem_used_percent

  • disk_used_percent

📸 Screenshot: CWAgent metrics showing live system data

---

Step 4: Create a CPU Utilization Alarm

Update your alarm-cpu.json with the EC2 instance ID:

"Value": "i-xxxxxxxxxxxxxxxxx"

Then run:

aws cloudwatch put-metric-alarm --cli-input-json file://alarm-cpu.json

📸 Screenshot: Alarm creation and active state

---

Step 5: Build a Custom Dashboard

• Go to CloudWatch → Dashboards → Create new

• Add widgets:

  • Line chart: CPU % over time

  • Gauge: Memory used %

  • Text widget: alarm state

📸 Screenshot: Dashboard with live metrics

---

💡 Optional Enhancements

• Add SNS notification to the alarm (email/slack)

• Stream logs to CloudWatch Logs for auditing

• Export the whole setup as CloudFormation or Terraform

---

📂 GitHub Repository

🔗 github.com/PradeepMahadevaiah/EC2_CloudWatch_Monitoring

Includes:

• cloudwatch-agent-config.json

• cloudwatch-setup.sh

• alarm-cpu.json

• README with full instructions

---

🧠 Key Learnings

✅ Hands-on with CloudWatch Agent configuration and metrics
✅ Created actionable alarms and visual dashboards
✅ Practiced AWS CLI for metric and alarm management

---

🙌 Conclusion

Monitoring is a critical DevOps skill — and CloudWatch makes it fully native to AWS. This project helped me solidify my understanding of infrastructure health, real-time metrics, and operational visibility.

Let me know your thoughts or feedback — or fork the repo to try it yourself!

---