AWS Café Application Optimization Project

Project Overview

I recently completed an optimization project for the Café web application on AWS. This activity required me to optimize the AWS resources used to run the Café web application. Specifically, I had to:

Uninstall the decommissioned local database from the Café instance to decrease the instance's storage requirements.
Change the instance type to T3 micro to reduce costs.

After completing this activity, I was able to:

Optimize an Amazon Elastic Compute Cloud (Amazon EC2) instance to reduce costs.
Use the AWS Pricing Calculator to estimate AWS service costs.

The business case for this was a new requirement for Café – to optimize resources to reduce AWS service costs.

After the migration to Amazon Relational Database Service (Amazon RDS) was completed (an action taken in a prior activity), I identified optimization opportunities to implement to reduce AWS service costs. First, I realized that the decommissioned local database could be uninstalled to reduce the storage requirements of the Café instance. I also realized that the Café instance could be downsized to a smaller instance type because the database process was no longer running on it.

Task 1: Optimize the website to reduce costs

Because the local database was migrated to Amazon RDS, I could reduce AWS service costs by performing the following actions on the Café EC2 instance:

Remove the local database from the instance. This action reduced costs in both CPU and storage utilization.
Change the instance type from t3.small to t3.micro. Because the database process no longer runs on the instance, the smaller instance type was both effective and also cheaper to run.

In this task, I used the AWS Command Line Interface (AWS CLI) to perform these actions. I began by opening a Secure Shell (SSH) session to the Café instance and the CLI Host.

Task 1.1.3: Configure the AWS CLI

Before I could run AWS CLI commands on the instance, I first had to configure the AWS CLI environment to define the AWS account credentials, Region name, and output format to use.

I discovered the region in which the CLI Host instance was running:
curl http://169.254.169.254/latest/dynamic/instance-identity/document | grep region
Then I updated the AWS CLI software with the credentials:
aws configure
At the prompts, I entered the following information:
- AWS Access Key ID: I clicked on the Details drop down menu above the instructions, and then clicked Show. I copied the AccessKey value and pasted it into the terminal window.
- AWS Secret Access Key: I copied and pasted the SecretKey value from the same Credentials screen.
- Default region name: I typed in the name of the region where my EC2 instances were running, which I discovered was (us-west-2).
- Default output format: json

I left this terminal window SSH session open to use it later in the activity.

Task 1.2: Connect to the CLI Host instance by using SSH

I followed the same instructions that I used in Task 1.1 to open an SSH session to a different EC2 instance—the CLI Host instance.
I did not close the connection to the CafeInstance, instead, I created a connection to the CLI Host in a new window (using putty on Windows or can be done using an additional terminal window on macOS/Linux).
I found the CLI Host public IP address in the EC2 console, and by clicking on the Details drop down menu above the instructions, and then clicking Show.
After connecting, I also configured the AWS CLI client software on the CLI Host EC2 instance, by running the aws configure command.

Task 1.3: Uninstall MariaDB and resize the instance

I stopped the local database and uninstalled it from the Café instance. In the SSH window for the CafeInstance, I entered:
sudo systemctl stop mariadb sudo yum -y remove mariadb-server
When the last command ran successfully, I saw a Complete! message in the output.
I closed the SSH window for the CafeInstance because I no longer needed it.
I determined the Instance ID of the CafeInstance. I switched to the SSH window for the CLI Host instance and entered:
aws ec2 describe-instances \ --filters "Name=tag:Name,Values= CafeInstance" \ --query "Reservations[*].Instances[*].InstanceId"
I recorded the value returned as:
CafeInstance Instance ID: i-nnnnnnnnnn
I stopped the Café instance and changed its instance type to t3.micro. In the SSH window for the CLI Host instance, I entered:
aws ec2 stop-instances --instance-ids <CafeInstance Instance ID>
In the command, I substituted <CafeInstance Instance ID> with the value that I recorded earlier.
I changed the instance type to t3.micro. In the SSH window for the CLI Host instance, I entered:
aws ec2 modify-instance-attribute \ --instance-id <CafeInstance Instance ID> \ --instance-type "{\"Value\": \"t3.micro\"}"
In the command, I substituted <CafeInstance Instance ID> with the value that I recorded earlier.
When the command completed successfully, no output was returned.
I started the Café instance. In the SSH window for the CLI Host instance, I entered:
aws ec2 start-instances --instance-ids <CafeInstance Instance ID>
In the command, I substituted <CafeInstance Instance ID> with the value that I recorded earlier.
I checked the current state of the instance, and waited until the status showed running. In the SSH window for the CLI Host instance, I entered:
aws ec2 describe-instances \ --instance-ids <CafeInstance Instance ID> \ --query "Reservations[*].Instances[*].[InstanceType,PublicDnsName,PublicIpAddress,State.Name]"
In the command, I substituted <CafeInstance Instance ID> with the value that I recorded earlier.
The instance took a few moments to reach the running state. I periodically repeated the command until I could confirm that it was running. I also recorded the PublicDnsName and PublicIPAddress values that were returned by the command by using the following format:
Downsized CafeInstance Public DNS Name: ec2-zzz-zzz-zzz-zzz.eu-west-2.compute.amazonaws.com Downsized CafeInstance Public IP Address: nnn.nnn.nnn.nnn

Information: Because I restarted the instance, Amazon EC2 assigned a different Public DNS name and Public IP address to the instance than what it had before.

I tested the Café website to make sure that it was functional. In a browser window, I entered the following URL:
http://<Downsized CafeInstance Public DNS Name>/cafe
I substituted <Downsized CafeInstance Public DNS Name> with the value that I recorded.
I exercised the website's functions to verify that it worked properly.

I successfully uninstalled the decommissioned local database and downsized the Café instance.

Task 2: Use the AWS Pricing Calculator to estimate AWS service costs

AWS provides a tool that allows users to estimate the monthly costs of the AWS services that they use or are planning to use. In this task, I used the AWS Pricing Calculator to estimate the cost of running the Café website on AWS before and after EC2 instance optimization. I then calculated the projected cost savings.

NOTE: The values that I entered into the AWS Pricing Calculator were simplified to serve the purposes of this exercise. The intent was to show the basic use of the calculator and highlight the functions that it provides.

The pricing examples shown in this activity were current as of time of publishing, April 2020, and is for informational purposes. Refer to the AWS website for current pricing by service.

Task 2.1: Calculate the costs before optimization

First, I calculated the costs of running the website in its before optimization topology, that is, on a T3 small instance with a decommissioned local database still occupying storage space.

Specifically, I used the following service list and configuration to describe the topology components:

Region: (the region where the CafeInstance EC2 instance is running)

Amazon EC2 instance:

Instance type: t3.small
Instance class: On-Demand
Utilization: 100% per month
Operating system: Linux
Amazon EBS volume: General Purpose SSD (gp2), 40 GB (including 20 GB occupied by the local database)

Amazon RDS instance:

Instance class: db.t3.micro
Engine: MariaDB
Allocated storage: 20 GB

I opened the AWS Pricing Calculator. In a web browser, I went to:
https://calculator.aws
I clicked on Create estimate.
I browsed down and chose Configure in the Amazon EC2 service box.
In the Region menu at the top of the page, I selected the region where the CafeInstance EC2 instance was running. For example, I chose US East (N. Virginia) if my instances were running in us-east-1.
When I was prompted to confirm the region change, I chose Change Region.
I chose the Advanced estimate option.
In the EC2 instance specifications area, for the Operating system, I chose Linux
In the Workload area:
- I chose Constant usage.
- For Number of instances I chose 1.
In the EC2 instances area, in the search box, I searched for and then selected the t3.small instance type.
In the Pricing strategy area, I set the pricing model to On-Demand
In the Amazon Elastic Block Storage (EBS) area:
- Storage for each EC2 instance: General Purpose SSD (gp2)
- Storage amount: 40 GB
- Snapshot Frequency: No snapshot storage
I scrolled to the bottom and selected Add to my estimate.

I had now estimated the cost of the EC2 instance.

Next, I added the RDS instance to my price estimate.

In the My Estimate page, I clicked Add service.
In the Select service page, I located and clicked Configure in the Amazon RDS for MariaDB service panel.
I configured as follows:
- Region: (chose the same Region I chose for the EC2 instance)
- MariaDB instance specifications: Standard (single-AZ)
- Instance type: searched for and selected db.t3.micro
- Quantity: 1
- Pricing model: On-Demand Instances
- Storage volume: General Purpose SSD (gp2)
- Storage amount: 20 GB per month
I chose Add to my estimate.
The My Estimate page showed a breakdown of the estimated monthly cost of the AWS services that I configured, and it provided a monthly total.
I chose Save and share.
When prompted, I chose Agree and continue.
I chose Copy the public link and pasted the link into another browser tab.
I exported the estimate to a comma-separated values (CSV) file by choosing Action > Export estimate.
In the export dialog window, I clicked OK and saved the file to my local computer. I opened it to see its contents.

Task 2.2: Calculate the costs after optimization

Next, I calculated the costs of running the website after the Café instance was optimized. Specifically, I modified the following entries in the calculator to reflect the effects of the optimization:

Amazon EC2 instance:

Instance type: t3.micro (Reduced size)
Amazon Elastic Block Store (Amazon EBS) General Purpose SSD (gp2), 20 GB (Reduced from 40 GB because the local database was removed)

In the https://calculator.aws/#/estimate browser tab, I clicked Edit next to the Amazon EC2 entry.
In the EC2 Instances area, I searched for and selected t3.micro as the instance type.
In the Amazon Elastic Block Storage (EBS) area, I changed the Storage amount to 20 GB.
I scrolled down and clicked Save to see the monthly cost estimate.
I exported the estimate to a comma-separated values (CSV) file by choosing Action > Export estimate.
In the export dialog window, I clicked OK and saved the file to my local computer. I opened it to see its contents.

Task 2.3: Estimate the projected cost savings for Café

Because I calculated the costs of the AWS services that are needed to run the Café website both before and after I optimized the instance, I could estimate the overall projected cost savings.

By removing the decommissioned local database and downsizing the Café instance type, I saved more in AWS service costs.

Martha and Frank were very happy that my initiative resulted in cost savings for the business.