You're reading from AWS for System Administrators Build, automate, and manage your infrastructure on the most popular cloud platform – AWS

Product type Paperback

Published in Feb 2021

Publisher Packt

ISBN-13 9781800201538

Length 388 pages

Edition 1st Edition

Tools

AWS

Concepts

System Administration

Author (1):

Prashant Lakhera

View More author details

Table of Contents (18) Chapters

Preface

1. Section 1: AWS Services and Tools

2. Chapter 1: Setting Up the AWS Environment FREE CHAPTER

3. Chapter 2: Protecting Your AWS Account Using IAM

4. Section 2: Building the Infrastructure

5. Chapter 3: Creating a Data Center in the Cloud Using VPC

6. Chapter 4: Scalable Compute Capacity in the Cloud via EC2

7. Section 3: Adding Scalability and Elasticity to the Infrastructure

8. Chapter 5: Increasing an Application's Fault Tolerance with Elastic Load Balancing

9. Chapter 6: Increasing Application Performance Using AWS Auto Scaling

10. Chapter 7: Creating a Relational Database in the Cloud using AWS Relational Database Service (RDS)

11. Section 4: The Monitoring, Metrics, and Backup Layers

12. Chapter 8: Monitoring AWS Services Using CloudWatch and SNS

13. Chapter 9: Centralizing Logs for Analysis

14. Chapter 10: Centralizing Cloud Backup Solution

15. Chapter 11: AWS Disaster Recovery Solutions

16. Chapter 12: AWS Tips and Tricks

17. Other Books You May Enjoy

Leave a review - let other readers know what you think

Custom CloudWatch metrics

If you look at the CloudWatch dashboard, there are four default metrics, as follows:

CPU
Disk I/O
Network
Instance/system status check

These default metrics (CPU, status check, and network) can be seen in the following screenshot:

Figure 12.3 – CloudWatch default metrics

But why are some of the standard metrics such as memory utilization or disk space not default metrics? The reason behind that is that an EC2 instance is a VM that emulates computer hardware such as CPU, random-access memory (RAM), and disk. Your AWS service can't look inside your instance because its operating system controls how many resources need to be allocated, such as how much memory is required. This is the main reason why it's not possible to determine memory utilization by looking at the virtual hardware. We need to install solutions such as CloudWatch agents to get these metrics and push them to CloudWatch...