Good news for developers, SREs, and cloud engineers — Amazon CloudWatch Agent now supports collecting detailed performance statistics for EBS volumes attached to EC2 and EKS nodes.

This means you can finally monitor and troubleshoot your EBS storage like a pro — with visibility into NVMe-level metrics such as:

• 🔁 IOPS (read/write operations)
• 📦 Throughput (bytes read/written)
• ⏱️ I/O wait time
• 🎯 Queue depth

Let’s break it down with a real-world example.

🔧 Use Case: App is Slow, But CPU & RAM Look Fine?

You’re running a production web app on EC2 with a gp3 EBS volume.
The app gets sluggish during peak hours, but CloudWatch shows:

• CPU: fine
• Memory: fine
• Network: fine

Now, thanks to the new update, you can collect EBS disk-level metrics and discover the real problem.

🧪 Step-by-Step Example

Step 1: Enable EBS Metrics in CloudWatch Agent

Update your amazon-cloudwatch-agent.json config:

{
  "metrics": {
    "metrics_collected": {
      "diskio": {
        "resources": ["*"],
        "measurement": [
          "reads", "writes", "read_bytes", "write_bytes",
          "io_time", "await", "util", "queue"
        ],
        "metrics_collection_interval": 60
      }
    }
  }
}

Then restart the agent:

sudo /opt/aws/amazon-cloudwatch-agent/bin/amazon-cloudwatch-agent-ctl \
  -a fetch-config -m ec2 -c file:/path/to/config.json -s

Step 2: View in CloudWatch

You'll now see custom metrics like:

• await → time the app waits for I/O
• queue → how many I/O ops are waiting
• io_time → total time EBS spends on operations
• read_bytes, write_bytes → data throughput

Step 3: Analyze & Act

During peak load:

• queue = 22 (too high)
• await = 120ms (delays noticeable)
• write_bytes drops sharply

🧠 Root cause: EBS is bottlenecked. Time to provision more IOPS or switch from gp3 to io2.

✅ Why This Matters

🧾 TL;DR

🚀 CloudWatch Agent now supports:

• NVMe-based EBS performance metrics
• Queue depth, IOPS, throughput, and more
• Alarms, dashboards, and smarter diagnostics

No more guessing — now you can see and solve storage bottlenecks confidently.

Have you started using EBS metrics in your monitoring stack?
Drop your setup or questions in the comments