Aisalkyn Aidarova

Posted on Oct 6, 2025

AWS introduction

#aws #beginners #tutorial

🧾 AWS Billing, Budget, and Cost Management – Step-by-Step

1. Access the Billing Console

In the AWS Management Console, click your account name (top right).
Choose Billing and Cost Management.

⚠️ If you see “Access Denied” — you are using an IAM user.
Only the root account can enable billing access for IAM users.

2. Enable Billing Access for IAM Users

Sign in as the root account owner.
Go to Account → IAM user and role access to billing information.
Turn ON “Activate IAM access.”

Now IAM users with admin permissions can view billing.

3. View Billing Data

Go back to Billing → Bills.
You can now see:
- Month-to-date costs
- Forecasted cost
- Charges by service (e.g., EC2, S3, etc.)
Click on a month → scroll to Charges by Service to identify what costs money (e.g., NAT Gateway, EBS, etc.).

4. Check the Free Tier

Go to Billing → Free Tier.
Shows:
- Usage vs Free Tier limit
- Forecasted overages
If usage turns red, you’re about to be charged — stop or delete resources.

5. Create a Budget

Go to Billing → Budgets → Create budget.
Choose “Use a template (simplified)”.

Option A: Zero Spend Budget

Template: Zero spend
Name: My Zero Spend Budget
Alert email: your address (e.g., yourname@example.com)
Sends alert as soon as you spend $0.01.

Option B: Monthly Cost Budget

Template: Monthly cost budget
Limit: e.g., $10
Add alert recipients.
Alerts at:
- 85% of actual spend
- 100% of actual spend
- 100% of forecasted spend

✅ Result

You’ll get an email warning if:

You spend 1 cent (zero spend budget), or
Your forecasted/actual spend hits your limit.

This ensures students don’t overspend during AWS labs.

☁️ Amazon EC2 – Elastic Compute Cloud

1. What is EC2?

EC2 = Elastic Compute Cloud
It’s AWS’s main Infrastructure as a Service (IaaS) offering.
Lets you rent virtual machines (VMs) — called EC2 Instances — on demand.
Foundation of AWS: most AWS services depend on EC2 behind the scenes.

2. Key EC2 Components

Component	Description
EC2 Instance	The virtual server you rent from AWS.
EBS Volume	Elastic Block Storage – a network-attached disk for your instance.
Elastic Load Balancer (ELB)	Distributes incoming traffic across multiple instances.
Auto Scaling Group (ASG)	Automatically increases or decreases the number of running instances based on demand.
Security Group	Acts as a firewall — controls inbound/outbound traffic to the instance.
Elastic IP	A static public IP address you can attach to your instance.
User Data	Script that runs once when the instance boots — used for automation/setup tasks.

3. Choosing EC2 Instance Settings

When launching an EC2 instance, you choose:

Option	Examples / Details
Operating System	Linux (most popular), Windows, or macOS.
Compute (vCPUs)	Choose instance type (e.g., `t2.micro`, `t3.medium`) based on performance.
Memory (RAM)	Depends on workload size (web server vs. database).
Storage	- EBS: network-attached, persistent storage. - Instance Store: local hardware disk (temporary).
Network	Select subnet, VPC, and network interface (speed, public IP, etc.).
Firewall Rules	Configure Security Groups — open only necessary ports (e.g., 22 for SSH, 80 for HTTP).

4. Bootstrapping with User Data

Bootstrapping = running setup commands automatically when the instance launches.

✅ Common tasks in User Data:

Update packages (yum update -y or apt update -y)
Install software (e.g., Nginx, Apache, Python)
Download configuration files
Start services automatically

🧠 Notes:

Runs only once at first boot.
Executed as root user (no need for sudo).
Makes EC2 setup automated and repeatable.

Example:

#!/bin/bash
yum update -y
yum install -y nginx
systemctl start nginx
systemctl enable nginx
echo "<h1>Hello from EC2</h1>" > /usr/share/nginx/html/index.html

5. Why EC2 Matters

Core building block of the AWS ecosystem.
Lets you quickly deploy servers on demand.
Forms the base for many other services (ECS, EKS, Beanstalk, etc.).
Teaches the foundation of cloud computing: scalability, pay-as-you-go, and automation.

🚀 Launching Your First EC2 Instance (Amazon Linux)

1. What You’ll Do

You will:

Launch your first EC2 instance (a virtual server).
Use User Data to automatically install a web server.
Access the website through a browser.
Learn to start, stop, and terminate the instance.

2. Launch an Instance

Step 1: Open EC2 Console

Go to AWS Management Console → EC2 → Instances
Click Launch Instances

Step 2: Name and Tags

Name: My First Instance
(Tag Key = Name, Value = My First Instance)

Step 3: Choose an AMI (Amazon Machine Image)

Go to Quick Start → Amazon Linux 2 AMI (64-bit x86)
✅ Free Tier eligible

This defines the operating system for your EC2 instance.

Step 4: Choose Instance Type

Choose t2.micro (Free Tier eligible)
1 vCPU, 1 GB RAM
Perfect for small practice servers

Step 5: Create or Choose Key Pair

You’ll need a key to connect via SSH later.

Name: EC2Tutorial
Type: RSA
Format:
- .pem → for Mac, Linux, or Windows 10+
- .ppk → for older Windows (PuTTY)
Download and save it safely — AWS will not let you download again!

Step 6: Configure Network Settings

Leave defaults (public IP assigned automatically).
Create a Security Group (default name: launch-wizard-1).
Add inbound rules:
- SSH (port 22) → Source: Anywhere
- HTTP (port 80) → Source: Anywhere (This allows browser access.)

Step 7: Configure Storage

Default: 8 GB gp2 EBS volume
You get up to 30 GB free under Free Tier.
Option “Delete on Termination” = Yes (keeps cleanup simple).

Step 8: Add User Data Script

Scroll to Advanced details → User data
Paste this script:

#!/bin/bash
yum update -y
yum install -y httpd
systemctl start httpd
systemctl enable httpd
echo "<h1>Hello World from $(hostname -f)</h1>" > /var/www/html/index.html

This will:

Update the system
Install Apache web server
Enable it at boot
Create a simple “Hello World” web page

Step 9: Launch

Review → Launch instance
Go to View all instances
Wait until Instance State = running (takes ~10–15 seconds)

3. Access the Website

Copy your Public IPv4 address
In your browser, enter:

  http://<Public-IP>

(⚠️ Use http, not https.)
You’ll see: Hello World from 172.x.x.x — where the number is the private IP.

4. Manage the Instance

Action	What It Does	Note
Stop Instance	Shuts down the server	You’re not billed while stopped
Start Instance	Restarts it later	⚠️ May get a new public IP
Terminate Instance	Deletes server + volume	Irreversible — removes data

🧠 The private IP stays the same, but the public IP changes after every stop/start cycle unless you assign an Elastic IP.

5. Key Takeaways

You can launch a web server in minutes without owning hardware.
User Data automates setup during boot.
Learn to stop/terminate to avoid charges.
EC2 is the core of cloud computing — flexible, fast, pay-as-you-go.

⚙️ Amazon EC2 Instance Types

1. Why EC2 Instance Types Exist

AWS offers different EC2 instance types to match different workloads —
from lightweight web servers to machine learning and databases.

Each type has:

Different CPU, RAM, and network capabilities
Different optimization (compute, memory, storage, or networking)
Different pricing

2. EC2 Instance Naming Convention

Example: m5.2xlarge

Part	Meaning	Example
m	Instance class / family	`m` = general purpose
5	Generation	`5` = newer than `4`, older than `6`
2xlarge	Size	Larger = more vCPUs & memory

So m5.2xlarge means:
➡️ General-purpose instance, generation 5, with 2xlarge size (moderate CPU and memory).

3. EC2 Instance Families (Main Categories)

Type	Family Prefix	Description	Example Use Cases
🧩 General Purpose	`t`, `m`, `a`	Balanced CPU, memory, and networking	Web servers, code repos, dev/test environments
⚡ Compute Optimized	`c`	High CPU performance	Batch processing, media encoding, gaming, ML inference
🧠 Memory Optimized	`r`, `x`, `z`	High RAM for in-memory processing	Databases, caching (Redis), analytics, BI
💾 Storage Optimized	`i`, `d`, `h`	High local disk throughput	Big data, OLTP, NoSQL, data warehousing
🎮 Accelerated Computing	`p`, `g`, `f`	GPUs or FPGAs	AI/ML training, deep learning, rendering, HPC

4. Common Instance Examples

Instance	vCPU	Memory (GB)	Optimized For
`t2.micro`	1	1	General-purpose, free tier
`m5.large`	2	8	Balanced web/app server
`c5.4xlarge`	16	32	Compute-intensive tasks
`r5.16xlarge`	64	512	Memory-heavy databases
`i3.8xlarge`	32	244	Storage-optimized, high IOPS

5. Helpful Reference Websites

🔗 AWS Official Instance Types Page

https://aws.amazon.com/ec2/instance-types
→ Lists all current instance families, pricing, and features.

🔗 EC2Instances.info

https://ec2instances.info
→ Excellent for:

Comparing CPU, memory, storage, and cost
Searching & filtering instance families
Quickly checking On-Demand and Reserved pricing

6. Key Takeaways

✅ Understand the prefixes:

t → test & dev (burstable)
m → general-purpose
c → compute-heavy
r → memory-heavy
i → storage-heavy
p/g → GPU-based

✅ Choose instance type by workload:

Web apps → t2.micro, m5.large
Databases → r5.xlarge
Machine learning → p3, g4
Data warehousing → i3, d2

✅ Use the AWS Free Tier:

t2.micro or t3.micro = Free for 12 months (750 hours/month)

🔒 Security Groups in Amazon EC2 (Firewalls)

1. What Are Security Groups?

Security Groups (SGs) are virtual firewalls that control traffic into and out of your EC2 instances.
They define network access rules based on:
- Ports (e.g., 22 for SSH, 80 for HTTP)
- Protocols (TCP, UDP, etc.)
- Source/Destination (IP ranges or other security groups)

🧠 Think of a security group as a “protective shield” that decides who can talk to your server and on which ports.

2. Basic Behavior

Direction	Default Behavior	Purpose
Inbound	❌ All traffic blocked by default	Protects the instance from unwanted access
Outbound	✅ All traffic allowed by default	Lets the instance connect to the internet (updates, downloads, etc.)

SGs contain only ALLOW rules (no explicit “deny”).
If traffic is not explicitly allowed → it’s implicitly denied.

3. How Security Groups Work

Example:

You (your computer) are on the public internet, trying to connect to an EC2 instance.

The EC2 instance has a Security Group attached.
That SG checks its inbound rules:
- If your IP and port (e.g., 22 or 80) match → traffic allowed
- If not → traffic blocked

Two traffic directions:

Inbound rules → from the outside → into EC2
Outbound rules → from EC2 → out to the internet

4. Security Group Rules Format

Field	Description	Example
Type	What kind of connection	SSH, HTTP, HTTPS
Protocol	Usually TCP	TCP
Port Range	Communication port	22, 80, 443
Source/Destination	IP or Security Group	`0.0.0.0/0` (all) or your IP

Example Rule:

Type	Protocol	Port	Source
SSH	TCP	22	Your IP (`203.x.x.x/32`)
HTTP	TCP	80	`0.0.0.0/0`

5. Key Characteristics

✅ You can attach:

One SG → multiple instances
One instance → multiple SGs

✅ SGs are:

Region-specific
VPC-specific

✅ SGs live outside the instance (so blocked traffic never reaches it).

✅ If your app times out, it’s probably an SG issue.
If you get connection refused, SG worked but the app isn’t running.

6. Best Practices

Create a dedicated SG for SSH (port 22) and restrict it to your IP only. Example: MySSH-SG → Inbound rule: SSH (22) → Source: your IP.
Create separate SGs for each application/service (e.g., web, database).
Regularly review inbound rules — remove unused ones.

7. Referencing Other Security Groups

Security groups can reference other security groups instead of IPs.

Why use it?

When instances must communicate internally (e.g., web server → database),
you don’t have to manage IP addresses.

Example:

SG-Web → allows inbound HTTP (80) from 0.0.0.0/0
SG-DB → allows inbound MySQL (3306) from SG-Web

➡️ Any instance with SG-Web can talk to instances with SG-DB over port 3306.

This is common with load balancers and multi-tier apps.

8. Common Ports to Remember

Port	Protocol	Purpose
22	SSH	Linux remote login
21	FTP	File Transfer Protocol
22	SFTP	Secure File Transfer (uses SSH)
80	HTTP	Unsecured web traffic
443	HTTPS	Secured web traffic
3389	RDP	Remote Desktop (Windows)

9. Quick Recap

✅ Inbound = Blocked by default
✅ Outbound = Allowed by default
✅ Security Groups = ALLOW rules only
✅ Region + VPC bound
✅ Timeout → SG issue, Connection Refused → App issue

🔐 Hands-On: Working with Security Groups in EC2

1. Where to Find Security Groups

In the EC2 Console, select your instance → click Security tab.
You’ll see:
- Inbound rules
- Outbound rules
- Linked Security Group(s)

For a full view:
👉 Left menu → Network & Security → Security Groups

2. Default Security Groups

You’ll typically see:

Default security group (created automatically per VPC)
Launch-wizard-1 (created during your first EC2 launch)

Each SG has:

A unique ID (e.g., sg-0a12b3c4d5e6f)
Inbound rules → traffic into your instance
Outbound rules → traffic out of your instance

3. Viewing and Editing Inbound Rules

Example: Launch-wizard-1

Type	Protocol	Port	Source	Purpose
SSH	TCP	22	0.0.0.0/0	Remote terminal access
HTTP	TCP	80	0.0.0.0/0	Web server access

These rules allowed us to:
- SSH into the instance (port 22)
- Access the web page (http://<Public-IP>) on port 80

4. Testing Firewall Behavior

🔸 Case 1 – Remove the HTTP rule

Delete the inbound rule for port 80.
Save changes.
Try reloading your website → ❌ Timeout

🧠 Timeout = Security Group issue

Your request never reached the EC2 instance.

Fix: check inbound rules.

🔸 Case 2 – Add the HTTP rule back

Add inbound rule:
- Type: HTTP
- Port: 80
- Source: Anywhere (0.0.0.0/0)
Save rules → ✅ Refresh your page → Works again!

The port 80 rule allows public HTTP access to your web server.

5. Adding New Rules

You can:

Choose any port or port range (e.g., 443 for HTTPS).
Pick from the dropdown list (common protocols).
Specify source:
- Anywhere (0.0.0.0/0) → open to everyone
- My IP → restrict access to your own machine
- Custom CIDR, security group, or prefix list → for advanced setups

⚠️ Note: If your IP changes (e.g., new Wi-Fi or VPN), you’ll lose access if rule is “My IP”.

6. Outbound Rules

Default outbound rule:

Type	Protocol	Port	Destination
All traffic	All	All	0.0.0.0/0

→ This allows your instance to download updates, connect to APIs, or reach the internet freely.

7. Multiple Security Groups and Instances

One EC2 instance can have multiple SGs attached.
One SG can be attached to multiple EC2 instances.
Combined rules are additive — all allowed traffic from each SG is permitted.

8. Quick Diagnostic Tip

Symptom	Meaning	Fix
❌ Timeout	Blocked by SG (no inbound rule)	Add correct inbound rule
⚠️ Connection Refused	App/service not running on port	Start service inside EC2

✅ Summary

SGs control inbound and outbound traffic.
They’re stateful — return traffic is automatically allowed.
Timeouts = SG misconfiguration, not instance failure.
Use least privilege → only open required ports.
One SG = many instances; one instance = many SGs.

🧩 Connecting to Your EC2 Instance

1. Why We Need It

After launching an EC2 instance, the next step is to connect inside the server —
to install software, check logs, or perform maintenance.

To do this securely, AWS provides several connection methods, depending on your computer’s operating system.

2. The SSH Protocol

🔐 What is SSH?

SSH (Secure Shell) is a protocol that allows secure, encrypted remote access to Linux servers.

It lets you:

Run commands directly on your EC2 instance.
Manage software, configurations, and troubleshooting.
Transfer files securely (via SFTP or SCP).

3. Methods by Operating System

Platform	Recommended Method	Tool	Notes
Mac / Linux	SSH command-line	Built-in terminal	Use `ssh -i your-key.pem ec2-user@<Public-IP>`
Windows 10 / 11	SSH (built-in PowerShell)	Use `ssh` command	Works the same as on Mac/Linux
Windows 7 / 8 (or older)	PuTTY	Separate application	Convert `.pem` → `.ppk` file first
Any OS (browser-based)	EC2 Instance Connect	AWS Console → “Connect” → “EC2 Instance Connect”	Easiest, no software setup

4. 🧠 EC2 Instance Connect (Recommended for Beginners)

✅ Advantages:

Works on Mac, Linux, Windows — any browser.
No setup, no key conversion, no CLI required.
Uses your AWS credentials securely.
Best for quick testing and short sessions.

⚠️ Limitation:

Currently supports Amazon Linux 2 and Ubuntu instances only.
Not ideal for automation or long-term maintenance.

5. ⚙️ When to Use SSH

For advanced work, scripting, or automation.
When using custom Linux distributions.
When setting up multiple servers with consistent access.

Example SSH Command (Mac/Linux/Win10+)

ssh -i ~/Downloads/ec2tutorial.pem ec2-user@<Public-IP>

Replace <Public-IP> with your instance’s address.
Ensure port 22 is open in your Security Group.

6. Common SSH Connection Issues

Problem	Likely Cause	Fix
❌ Timeout	Security Group missing port 22 rule	Add inbound rule for SSH (port 22, your IP)
⚠️ Permission denied (publickey)	Wrong key or wrong user name	Use correct `.pem` and correct user (`ec2-user` for Amazon Linux)
⛔️ Connection refused	Instance not running or booting	Wait for “Running” state
🌐 Wrong IP	Instance stopped/restarted	Use new Public IP or attach an Elastic IP

7. 💡 Instructor Tips

Only one method needs to work (SSH or EC2 Instance Connect).
Don’t stress if SSH fails — you’ll still progress fine with EC2 Instance Connect.
Keep your .pem key safe — AWS doesn’t allow redownloads.
Always check Security Group rules before troubleshooting deeper.

8. 🧭 Next Steps

Identify your OS.
Use the right connection method:

Mac/Linux → SSH
Windows 10+ → PowerShell SSH
Windows 7/8 → PuTTY
Any OS → EC2 Instance Connect
1. Connect and explore your EC2 server.

💻 Connecting to EC2 with SSH (Mac or Linux)

1. 🎯 Goal

Use SSH (Secure Shell) to remotely access your EC2 instance from your local terminal.
Once connected, you’ll be able to:

Run Linux commands directly on the EC2 machine
Verify network connectivity
Manage and troubleshoot your cloud server

2. 🧱 How SSH Works

Diagram:

Your Laptop (SSH Client)
   ↓ Port 22 (SSH)
Internet
   ↓
EC2 Instance (Amazon Linux 2)
Security Group → allows Port 22 inbound

Explanation:

SSH uses Port 22 to securely connect to the server.
The Security Group must allow inbound access on Port 22.
The connection authenticates using your private key (.pem).

3. 🧩 Preparation Steps

Locate your downloaded key file (e.g. EC2Tutorial.pem).

Rename it to remove spaces → ✅ EC2Tutorial.pem
1. Move it to a safe folder (e.g. ~/aws-course/).
2. In AWS Console:
Go to EC2 → Instances
Copy your Public IPv4 Address

Check Security Group → must allow:

 Type: SSH | Protocol: TCP | Port: 22 | Source: 0.0.0.0/0

4. 🖥️ Navigate to the Key File

Open your terminal:

cd ~/aws-course
ls

You should see:

EC2Tutorial.pem

If not:

Use pwd to see where you are.
Use cd .. to go up a directory until you find your folder.

5. 🔑 Set Proper Permissions

Your key file must not be publicly viewable:

chmod 400 EC2Tutorial.pem

This means: only you can read the file.

6. 🌐 Connect via SSH

Run:

ssh -i EC2Tutorial.pem ec2-user@<Public-IP>

Example:

ssh -i EC2Tutorial.pem ec2-user@54.165.90.11

-i → specify your private key file
ec2-user → default username for Amazon Linux 2
<Public-IP> → your EC2’s public IPv4 address

If prompted with:

Are you sure you want to continue connecting (yes/no)?

→ type yes.

7. ✅ You’re In!

If successful, your prompt changes:

[ec2-user@ip-172-31-45-20 ~]$

You’re now inside your EC2 instance.

8. 🧪 Try Basic Commands

whoami         # shows current user (ec2-user)
hostname       # displays machine name
ping google.com

Press Ctrl + C to stop the ping.

9. 🚪 Exit the SSH Session

To disconnect:

exit

or press Ctrl + D.

10. ⚠️ Important Notes

If you stop and start your instance → the Public IP changes. Update your SSH command accordingly.
Keep your .pem file secure — you cannot re-download it from AWS.
If you see "Permission denied", check:
- File permissions (chmod 400)
- Username (ec2-user)
- Correct IP address
- Port 22 open in security group

🧠 Summary

Step	Command	Purpose
1	`chmod 400 EC2Tutorial.pem`	Secure key permissions
2	`ssh -i EC2Tutorial.pem ec2-user@<Public-IP>`	Connect to instance
3	`whoami` / `ping google.com`	Test access
4	`exit`	Disconnect safely

💻 Connecting to EC2 Using SSH on Windows (PuTTY Method)

1. 🎯 Goal

Learn how to connect (SSH) from a Windows computer to an Amazon Linux 2 EC2 instance using PuTTY.

SSH lets you:

Control your EC2 instance remotely from Windows.
Run commands directly on your cloud server.
Troubleshoot or configure your Linux machine securely.

2. 🌍 How It Works

Your Windows PC (PuTTY)
   ↓ Port 22 (SSH)
Internet
   ↓
EC2 Instance (Amazon Linux 2)
Security Group → allows Port 22 inbound

✅ SSH (Secure Shell) runs over Port 22
✅ The EC2 Security Group must allow:

Type: SSH
Protocol: TCP
Port: 22
Source: 0.0.0.0/0

3. 🧩 Prerequisites

You already launched an EC2 instance (Amazon Linux 2).
You downloaded your key pair file (e.g., EC2Tutorial.pem).
You are using Windows 7, 8, or older (PuTTY required).

For Windows 10+, you can use PowerShell SSH instead.

4. 🧰 Install PuTTY Tools

Go to https://www.putty.org/
Download and install:

PuTTY (main SSH app)
PuTTYgen (key converter tool)

During setup → click Next → Install → Finish ✅

5. 🔑 Convert Your .PEM Key to .PPK

PuTTY requires .ppk format for private keys.

Steps:

Open PuTTYgen.
Click Load.
Navigate to your .pem file (e.g., EC2Tutorial.pem).

If it’s not visible → choose All Files (.) at the bottom right.
1. Select the file → click Open.
2. You’ll see: "Successfully imported foreign key"
3. Click Save private key → choose a name like EC2Tutorial.ppk.
4. When asked about a passphrase → click Yes (no passphrase needed).
5. Save it (e.g., on your Desktop).

✅ You now have both:

EC2Tutorial.pem  (AWS original)
EC2Tutorial.ppk  (PuTTY-compatible)

6. ⚙️ Configure PuTTY to Connect

Open PuTTY.
In Host Name (or IP address), enter:

   ec2-user@<Public-IP>

Example:

   ec2-user@54.167.123.45

Port: 22
Connection type: SSH
In the Category list → expand SSH → click Auth.
Under “Private key file for authentication” → click Browse.
Select your .ppk file (e.g., EC2Tutorial.ppk).
Go back to Session (top of the list).
Under “Saved Sessions,” name it something like:

   EC2-Instance

Then click Save.

✅ This stores your connection and key configuration.

7. 🔌 Connect to EC2

Select your saved session (EC2-Instance).
Click Open.
A security alert appears:

“The server’s host key is not cached in the registry.”
→ Click Yes (to trust it).

You’ll see:

   login as:

Type:

   ec2-user

✅ You’re in your EC2 instance!

8. 🧪 Test Commands

Inside PuTTY, try:

whoami          # shows current user
hostname        # shows machine name
ping google.com # tests internet connectivity

To stop the ping → Ctrl + C.

9. 🚪 Exit and Reconnect

To leave the session:

  exit

Next time:
- Open PuTTY
- Load your saved session (EC2-Instance)
- Click Open
- You’ll be logged in instantly — no need to reconfigure.

10. ⚠️ Common Troubleshooting

Problem	Likely Cause	Fix
❌ Timeout	Missing SSH rule in Security Group	Add inbound rule for port 22
⚠️ No auth methods available	Didn’t attach the `.ppk` key	Re-add the private key under SSH → Auth
⛔️ Permission denied (publickey)	Wrong username	Use `ec2-user` (not root)
🌐 Connection refused	Instance not running or wrong IP	Start instance and use new Public IP

🧠 Key Takeaways

PuTTY is the SSH tool for Windows 7/8.
Always convert .pem → .ppk using PuTTYgen.
Use ec2-user as the default login for Amazon Linux 2.
Always check port 22 is open in your Security Group.
Save your session — it saves time for future logins.

💻 Connecting to EC2 Using SSH on Windows 10 (PowerShell)

1. 🎯 Goal

Learn to connect (SSH) from a Windows 10 or later machine directly to your EC2 instance — without PuTTY — using PowerShell or Command Prompt.

2. 🔐 What Is SSH?

SSH (Secure Shell) lets you:

Remotely control your EC2 Linux server through a command line.
Run, install, or troubleshoot applications securely.
Avoid using any GUI — all actions happen via text commands.

3. 🌍 How It Works

Your Windows 10 PC (PowerShell SSH)
   ↓ Port 22 (SSH)
Internet
   ↓
EC2 Instance (Amazon Linux 2)
Security Group → allows Port 22 inbound

✅ Port 22 must be open in the Security Group:

Type: SSH | Protocol: TCP | Port Range: 22 | Source: 0.0.0.0/0

4. 🧩 Check If SSH Is Available

Open PowerShell or Command Prompt and type:

ssh

If you see command help (e.g. usage options) → SSH is installed. ✅
If not → install the Windows “OpenSSH Client” feature or use PuTTY (see previous lecture).

5. 📁 Locate Your Key File

Your key file is the .pem file you downloaded from AWS (e.g. EC2Tutorial.pem).

Steps:

Place it somewhere simple — e.g. Desktop or C:\Users\<YourName>\aws-course
In PowerShell:

   cd .\Desktop
   ls

You should see your .pem file listed.

6. ⚙️ Connect Using SSH

The command format is:

ssh -i "EC2Tutorial.pem" ec2-user@<Public-IP>

Example:

ssh -i "EC2Tutorial.pem" ec2-user@3.94.152.11

Explanation:

Part	Meaning
`-i`	Path to your private key (.pem file)
`ec2-user`	Default Linux username for Amazon Linux 2
`@<Public-IP>`	The public IPv4 address of your EC2 instance

7. ⚠️ First-Time Connection

You’ll see:

The authenticity of host ... can't be established.
Are you sure you want to continue connecting (yes/no)?

→ Type yes
✅ You are now inside your EC2 instance!

8. 🧰 If You Get Permission Errors

Windows sometimes restricts .pem file permissions, causing:

Permissions for 'EC2Tutorial.pem' are too open.

Fixing Permissions:

Right-click your .pem file → Properties
Go to Security tab → Advanced
Make sure:

Owner = your Windows user account
Click Change if needed → type your username → Check Names → OK
1. Click Disable inheritance → select Remove all inherited permissions
2. Click Add → Select a principal
Type your username → Check Names → OK
Give yourself Full control
1. Apply and close all dialogs.

✅ Now only you (the owner) have access to the key — SSH will work without warnings.

9. 🧪 Verify Connection

Once connected, try:

whoami          # shows the current user
hostname        # shows the EC2 machine name
ping google.com # tests internet connectivity

Press Ctrl + C to stop the ping.

10. 🚪 Exit the Session

To disconnect:

exit

or press Ctrl + D.

11. 🧠 Key Tips

Action	Command/Note
Check SSH installed	`ssh` in PowerShell or CMD
Navigate to key location	`cd .\Desktop`
Connect	`ssh -i EC2Tutorial.pem ec2-user@<Public-IP>`
Fix permissions	Change file owner + disable inheritance
Exit	`exit` or Ctrl + D
Public IP changes	If you stop/start EC2 → use new IP address

✅ Summary

Windows 10+ has built-in SSH — no need for PuTTY.
Use your .pem key from AWS to authenticate.
Adjust file permissions if Windows blocks access.
Always ensure port 22 is open in your Security Group.

🌐 Connecting to EC2 Using EC2 Instance Connect (Browser-Based SSH)

1. 🎯 Goal

Learn to connect to your Amazon EC2 instance directly from the AWS Console —
without needing any .pem key, PuTTY, or terminal setup.

This method works on Windows, Mac, and Linux using only a web browser.

2. ⚙️ What Is EC2 Instance Connect?

EC2 Instance Connect is a browser-based SSH client built into AWS.

It:

Lets you open a secure terminal session in your browser.
Uses a temporary SSH key (uploaded automatically by AWS).
Requires no manual key management.
Works with Amazon Linux 2 and Ubuntu instances.

3. 🧱 How It Works

Browser (AWS Console)
   ↓ HTTPS (Port 443)
AWS EC2 Instance Connect Service
   ↓ Temporary SSH key (Port 22)
EC2 Instance (Amazon Linux 2)
Security Group → must allow inbound port 22

✅ Behind the scenes, it still uses SSH —
so port 22 must be open in your Security Group.

4. 🚀 Step-by-Step: Connecting

Go to EC2 Console → Instances
Select your instance (e.g. My First Instance)
Click Connect (top right)
Choose EC2 Instance Connect (browser-based SSH)
Confirm details:

Instance ID: prefilled
Public IPv4 address: visible
Username: ec2-user (default for Amazon Linux 2)
1. Click Connect

✅ Within seconds, a new browser tab opens — you are now inside your EC2 instance terminal!

5. 🧪 Try Some Commands

In the browser terminal:

whoami          # shows 'ec2-user'
hostname        # shows internal hostname
ping google.com # test connectivity

Press Ctrl + C to stop the ping.

6. 🔒 Troubleshooting EC2 Instance Connect

❌ Connection Error:

“There was a problem connecting to your instance”

✅ Fix: Ensure Port 22 Is Open

Go to EC2 → Security Groups
Select the group attached to your instance.
Click Edit inbound rules
Add:

   Type: SSH | Protocol: TCP | Port Range: 22 | Source: 0.0.0.0/0

(Optional) Add IPv6 rule if needed:

   Type: SSH | Protocol: TCP | Port Range: 22 | Source: ::/0

Save changes → try connecting again.

7. 🧠 Key Points to Remember

EC2 Instance Connect is quickest for beginners — no setup, no key downloads.
It still relies on SSH port 22 — inbound rules must allow access.
It uses temporary credentials valid only for the session.
If you remove the SSH rule → connection fails immediately.
Works best for short admin sessions or training labs.

8. ✅ Summary

Feature	EC2 Instance Connect
Setup required	None (browser only)
SSH key needed	Temporary key handled by AWS
Port required	22 (SSH)
Supported OS	Amazon Linux 2, Ubuntu
Security	Uses HTTPS + ephemeral SSH key
Best for	Quick access, demos, and student labs

🔐 Using IAM Roles with EC2 Instances

1. 🎯 Goal

Learn how to securely give your EC2 instance permission to access AWS services using IAM Roles —
without storing Access Keys or running aws configure.

2. 💡 Why IAM Roles?

In AWS, EC2 instances often need to access other services (like S3, DynamoDB, or IAM).
You could use AWS credentials (Access Key + Secret Key) — but that’s unsafe.

❌ Bad Practice:

Running:

aws configure

and entering your personal IAM user credentials exposes them to anyone with instance access.
They could retrieve keys and use them elsewhere — a major security risk.

✅ Correct Practice:

Use an IAM Role attached to the EC2 instance.
AWS automatically injects temporary credentials through the instance metadata service.

3. ⚙️ How It Works

EC2 Instance
   ↕
IAM Role attached
   ↕
AWS automatically provides temporary credentials
   ↕
Access to AWS services (like IAM, S3, DynamoDB)

✔ No keys stored
✔ Rotates automatically
✔ Least privilege by policy

4. 🧪 Hands-On: Attach an IAM Role to EC2

Step 1: Connect to EC2

Use EC2 Instance Connect or SSH — both open a terminal inside your EC2.

In the shell, verify connection:

whoami
ping google.com

Then clear the screen:

clear

Step 2: Test AWS CLI Access

Try:

aws iam list-users

You’ll see:

Unable to locate credentials. You can configure credentials by running "aws configure".

This confirms your instance currently has no permissions.

Step 3: Create an IAM Role (if not already)

In AWS Console → IAM → Roles → Create role

Trusted entity: AWS Service
Use case: EC2
Attach permissions policy:

Choose IAMReadOnlyAccess (for demo)
1. Name: DemoRoleForEC2
2. Create role

Step 4: Attach Role to Instance

In AWS Console:

Go to EC2 → Instances
Select your instance → Actions → Security → Modify IAM Role
From the dropdown, choose DemoRoleForEC2
Click Save

Now go to the Security tab of your instance —
you’ll see:

IAM Role: DemoRoleForEC2

Step 5: Test Again

Back in the terminal:

aws iam list-users

✅ You now get a proper IAM response:

{
  "Users": [
    {
      "UserName": "AdminUser",
      "Arn": "arn:aws:iam::123456789012:user/AdminUser",
      ...
    }
  ]
}

Step 6: Remove and Re-Test

Detach the policy from your IAM role (in IAM console → Role → Permissions → Detach policy).

Then rerun:

aws iam list-users

❌ Now you get:

An error occurred (AccessDenied) when calling the ListUsers operation: User is not authorized to perform iam:ListUsers

✅ This proves that permissions are directly controlled by the IAM Role.

5. 🔁 How the Role Credentials Work

You can check the temporary credentials with:

curl http://169.254.169.254/latest/meta-data/iam/security-credentials/

This returns your IAM role name and a short-lived credential set.
AWS automatically rotates these keys for you.

6. 🧠 Key Takeaways

Concept	Explanation
IAM Role	Securely grants permissions to EC2 without credentials
Policy	Defines what actions EC2 can perform
AWS CLI	Automatically uses temporary role credentials
Never use `aws configure`	Don’t hardcode keys on EC2
IAMReadOnlyAccess	Safe policy to view IAM data
Metadata service	Provides auto-rotated credentials to EC2

✅ Summary

Never use static IAM keys inside EC2.
Always attach an IAM Role to the instance.
The role defines what AWS actions the instance can perform.
IAM credentials are automatically provided and rotated by AWS.
Use aws cli commands directly — no configuration needed.

💰 EC2 Instance Purchasing Options

1. 🎯 Goal

Understand how AWS offers different pricing models for EC2 —
each optimized for cost, flexibility, or stability — depending on workload type.

2. ⚙️ Overview

AWS gives you six main purchasing options for EC2 instances:

Type	Duration	Best For	Discount (vs On-Demand)	Reliability
On-Demand	Pay-as-you-go	Short, unpredictable workloads	–	⭐⭐⭐⭐⭐
Reserved Instances (RI)	1 or 3 years	Predictable, long-term workloads	Up to 72%	⭐⭐⭐⭐⭐
Savings Plans	1 or 3 years	Long-term spend commitment (flexible usage)	Up to 72%	⭐⭐⭐⭐⭐
Spot Instances	Variable	Short, flexible, fault-tolerant tasks	Up to 90%	⭐⭐
Dedicated Hosts / Instances	Long-term	Compliance, licensing, or isolation needs	– / Up to 70%	⭐⭐⭐⭐⭐
Capacity Reservations	Flexible	Reserved AZ capacity without discount	0%	⭐⭐⭐⭐

3. 🧩 1️⃣ On-Demand Instances

Pay per second (Linux/Windows) or per hour (other OS).
No upfront cost, no commitment.
Highest flexibility, highest price.
Perfect for:
- Testing, proof of concept, dev environments.
- Unpredictable workloads.

💡 Example:

“Run when you want, stop when you want — like renting a car by the hour.”

4. 💡 2️⃣ Reserved Instances (RI)

Commit for 1 or 3 years.
Save up to 72% compared to On-Demand.
Fixed attributes: instance type, region, tenancy, OS.
Payment options:
- No upfront
- Partial upfront
- All upfront (max discount)
Two types:
- Standard RI → fixed configuration
- Convertible RI → change family, OS, or size (discount ~66%)
Can buy/sell on AWS RI Marketplace.

✅ Use for:
Steady workloads like databases, web servers, or ERP systems.

5. 💸 3️⃣ Savings Plans

Modern alternative to RI.
Commit to spend a fixed $/hour (e.g., $10/hour for 3 years).
AWS automatically applies discount to matching compute usage.
Flexibility:
- Any instance size in same family.
- Switch between Linux ↔ Windows.
- Works with EC2, Fargate, Lambda.

✅ Use for:
Dynamic environments where workload changes but total spend is predictable.

💡 Analogy:

“You commit to spending $300 per month at a hotel — you can change rooms anytime.”

6. ⚡ 4️⃣ Spot Instances

Up to 90% cheaper than On-Demand.
AWS reclaims instances anytime with a 2-minute warning.
Ideal for workloads tolerant to interruptions:
- Batch jobs, rendering, data analysis, CI/CD runners, machine learning training.
Not suited for:
- Databases or critical systems.

💡 Analogy:

“Like last-minute hotel deals — super cheap, but you might get kicked out anytime.”

7. 🏠 5️⃣ Dedicated Hosts & Dedicated Instances

Dedicated Host

Physical server fully reserved for your account.
Visibility into underlying sockets, cores, VMs.
Use for:
- Bring Your Own License (BYOL) software (Oracle, SQL Server, etc.).
- Compliance or regulatory isolation needs.
Billed per host, can be reserved 1 or 3 years.

Dedicated Instance

Runs on hardware dedicated to you, but AWS manages placement.
You don’t see or control the physical server.
Slightly cheaper than Dedicated Host.

💡 Difference:

Feature	Dedicated Instance	Dedicated Host
Control over placement	❌	✅
Hardware visibility	❌	✅
Licensing (BYOL) support	❌	✅
Cost	Lower	Higher

8. 🧱 6️⃣ Capacity Reservations

Reserve capacity in a specific Availability Zone (AZ).
No discount — billed at On-Demand rate.
Guarantees instance availability even during high demand.
Can be canceled anytime.

✅ Use for:

Mission-critical workloads that must always launch.
Short-term events or DR (disaster recovery) readiness.

💡 Analogy:

“You book a hotel room but pay even if you don’t stay — you’re guaranteed it’s there.”

9. 🧮 Cost & Use Case Comparison

Option	Duration	Commitment	Discount	Suitable For	Risk of Interruption
On-Demand	None	None	–	Short, unpredictable workloads	❌
Reserved Instance	1–3 yrs	Fixed	Up to 72%	Steady usage	❌
Savings Plan	1–3 yrs	Spend commitment	Up to 72%	Flexible long-term	❌
Spot Instance	None	Variable	Up to 90%	Short, interruptible	✅
Dedicated Host	1–3 yrs	Fixed	–	Compliance, BYOL	❌
Capacity Reservation	Any	None	0%	Guaranteed capacity	❌

10. 🏨 Hotel Analogy (Easiest to Remember)

Option	Analogy	Description
On-Demand	Walk-in guest	Pay full price, come and go anytime
Reserved Instance	Long-term resident	Pay less for committing to stay longer
Savings Plan	Monthly membership	Spend fixed $ each month, flexible room type
Spot Instance	Last-minute deal	Cheap, but may lose your room anytime
Dedicated Host	Rent the whole hotel	Full control, private property
Capacity Reservation	Reserve a room just in case	Pay even if you don’t use it

11. 🧠 Exam & Interview Tips

❗On-Demand → short, unpredictable, no commitment.
💡Reserved Instance → predictable workloads (DBs, web apps).
💰Savings Plan → flexible workloads, commit to spend.
⚙️Spot → batch, ML, non-critical compute.
🧾Dedicated Host → compliance or BYOL licensing.
🧩Capacity Reservation → guaranteed AZ availability.

✅ Summary

Feature	Optimized For	Example
On-Demand	Flexibility	Dev/test, startups
Reserved	Predictability	Databases
Savings Plan	Spending control	Constant EC2 usage
Spot	Cost savings	Batch, analytics
Dedicated Host	Compliance & licenses	Oracle workloads
Capacity Reservation	Availability guarantee	Disaster recovery

⚡ Deep Dive: EC2 Spot Instances

1. 🎯 Goal

Learn how EC2 Spot Instances work, how to use them safely, and how AWS manages interruptions, pricing, and automation for massive cost savings.

2. 💰 Why Spot Instances?

Up to 90% cheaper than On-Demand.
You use unused EC2 capacity that AWS sells at a discount.
You must be prepared for interruptions.

✅ Ideal for:

Batch jobs
CI/CD runners
Data analytics
ML training
Image/video processing
Container clusters (ECS, EKS)

❌ Not ideal for:

Databases
Stateful apps
Long-lived sessions
Mission-critical production workloads

3. ⚙️ How Spot Pricing Works

🔸 Step 1 — You define:

Max Spot Price = the most you’re willing to pay/hour

🔸 Step 2 — AWS publishes:

Current Spot Price (varies by instance type & AZ)

🔸 Step 3 — If:

Condition	Result
Spot price ≤ Max price	✅ Instance runs
Spot price > Max price	⚠️ Instance interrupted (2-minute warning)

4. ⏰ Two-Minute Interruption Notice

When AWS reclaims your instance, you get a 2-minute warning.
You can:

Stop the instance → retain EBS data; restart later.
Terminate the instance → lose ephemeral data; cheaper.

💡 Choose based on workload type:

Stop → stateful compute
Terminate → stateless batch jobs

5. 🧱 Spot Blocks (Fixed-Duration Instances)

Lock a spot instance for 1–6 hours.
AWS guarantees no interruption during that period (except in rare capacity loss).
Cost is higher than standard Spot, but still cheaper than On-Demand.

✅ Best for predictable short jobs (e.g., nightly builds, simulations).

6. 📈 Spot Price Behavior

Prices fluctuate by Availability Zone and instance family.
Reflect supply and demand — not user bidding anymore (AWS sets the price).
Typically stable, but may spike if capacity tightens.

💡 Example:
m4.large On-Demand: $0.10/hr
Spot average: ~$0.04/hr → 60%+ savings

7. 🧩 Spot Requests

A Spot Request defines:

Number of instances
Max price
AMI, instance type, subnet
Duration (valid from/until)
Request type → One-Time or Persistent

🔹 One-Time Request

Launches once → fulfilled → ends automatically.
Good for single batch jobs.

🔹 Persistent Request

Stays open until canceled.
If an instance is terminated due to price/capacity, AWS automatically relaunches new ones when conditions improve.

8. ❌ How to Cancel Spot Requests Properly

⚠️ Order matters:

Cancel the Spot Request → Prevents AWS from launching replacements.
Terminate the Spot Instances → Frees resources you’re billed for.

If you terminate first (without canceling),
the Spot Request sees “0 instances running” and relaunches them again.

✅ Exam Tip → Always cancel request first, then terminate instances.

9. 🚀 Spot Fleets

A Spot Fleet = group of Spot + (optional) On-Demand instances that AWS manages to meet a target capacity at lowest possible cost.

📦 What You Define

Target capacity (e.g., 100 vCPUs or 10 instances)
Multiple launch pools:
- Different instance types
- Different AZs
- Different OSs
Allocation strategy (below)

⚙️ Allocation Strategies

Strategy	Description	Best For
Lowest-Price	Chooses the cheapest pool	Cost-optimized short workloads
Diversified	Spreads across multiple pools	Availability-focused, long workloads
Capacity-Optimized	Chooses pools with best capacity	Large-scale, reliable compute
Price-Capacity-Optimized	Balances lowest price + available capacity	🔹 Best for most real workloads

✅ AWS automatically replaces lost instances to maintain capacity.

10. 🔄 Spot Fleet vs Simple Spot Request

Feature	Spot Request	Spot Fleet
Single instance type	✅	❌
Multiple instance types	❌	✅
Across multiple AZs	❌	✅
Includes On-Demand	❌	✅
Auto-optimization	❌	✅
Best for	Simple job	Cost-optimized scaling cluster

💡 Think of Spot Fleet as an intelligent manager that keeps your compute capacity running at the lowest possible cost.

11. 💡 Practical Examples

Use Case	Best Approach
Hadoop/Spark batch jobs	Spot Fleet (diversified)
CI/CD pipelines	Spot Block (1–6 hrs)
ML model training	Spot Fleet (price-capacity-optimized)
Web servers with ASG	Combine On-Demand + Spot mix
Databases	Never use Spot

12. 🧠 Exam & Interview Tips

Spot = cheapest but interruptible.
2-minute warning before termination.
Spot Block = 1–6 hr fixed duration.
Cancel request → then terminate instances.
Spot Fleet optimizes across types, AZs, and prices.
Price-Capacity-Optimized = best modern default.
Don’t use Spot for critical or stateful systems.

✅ Summary

Feature	Description
Max discount	Up to 90%
Billing unit	Per second
Interruption notice	2 minutes
Typical use cases	Batch, analytics, CI/CD
Don’t use for	Databases, critical workloads
Key services	Spot Request, Spot Fleet
Recommended strategy	Price-Capacity-Optimized

🚀 All the Ways to Launch EC2 Instances

1. 🎯 Goal

Understand every method AWS offers to launch EC2 instances, from Spot Requests to Dedicated Hosts, and when each is appropriate for cost, flexibility, or compliance.

2. ⚡ Option 1 — Spot Requests

💰 Save up to 90% on compute costs!

A Spot Request asks AWS for spare EC2 capacity at discounted pricing.

🧭 Steps in the Console

In the EC2 Dashboard, go to Spot Requests.
Click Pricing history → view past 3 months for any instance type (e.g., c4.large).

Black bar = On-Demand price.
Colored lines = Spot prices per AZ.
Typically 60–70% cheaper and quite stable.

🧱 Create a Spot Request

Click Request Spot Instances → You can either:

Use a Launch Template, or
Manually configure launch settings:
- AMI (e.g., Amazon Linux 2)
- Key pair
- VPC/subnet
- Security group

⚙️ Request Details

Setting	Description
Max Price	Max hourly rate you’re willing to pay. If AWS’s Spot price rises above this, the instance is stopped or terminated.
Valid From / Valid Until	Defines the active time window of your request.
Terminate when expired	Decide whether to stop instances when request expires.
Load Balancer/Target Group	(Optional) attach to ELB/ALB target group.

📊 Target Capacity

Define how many instances or vCPUs you want.
Choose to maintain capacity — AWS will automatically re-launch if any are lost.
Interruption behavior: terminate, stop, or hibernate.

🌐 Networking

Choose VPC, subnet, and Availability Zone (AZ).
Pick instance types manually (e.g., c3.large, c4.large) or define attribute filters:
- Min/max vCPUs
- Min/max memory
- Architecture, virtualization type, etc.

💡 The broader your filters → the more flexibility → the cheaper AWS can provide capacity.

🧮 Allocation Strategy

Strategy	Description	Use Case
Lowest Price	Choose pools with lowest Spot price	Short workloads, max savings
Capacity Optimized	Prefer pools with highest capacity	Large workloads
Diversified	Spread across multiple pools	High availability
Price-Capacity Optimized	Mix of cost + reliability	✅ Recommended for most users

You can also maintain a diverse pool of instance types for resilience.

🧾 Example

Target capacity: 10 instances
Estimated fleet cost: $0.156/hr
Savings: ~73% vs On-Demand

3. ⚡ Option 2 — Launching Spot Instances Directly

Instead of Spot Fleet, you can launch directly from:
EC2 → Instances → Launch Instance → Advanced details → Request Spot Instances

You’ll see options:

Request type: one-time (default) or persistent
Max price: default = On-Demand price (can customize)
Interruption behavior: stop, terminate, hibernate
Request validity: specify start & end time

💡 The “block duration” (1–6 hour Spot blocks) feature was deprecated after Dec 2022.

4. 💸 Option 3 — Reserved Instances (RI)

Buy capacity in advance for 1 or 3 years.

Console Flow:

EC2 → Reserved Instances
Search instance type (e.g., c5.large)
Choose:

Term: 12 or 36 months
Type: Standard or Convertible
Payment: All Upfront / Partial / No Upfront
1. Add to Cart → View Cart → (⚠ don’t actually purchase unless needed!)

💡 Convertible RIs let you change instance family/OS.
Standard RIs are locked but cheaper.

⚠️ Note: RIs are slowly being replaced by Savings Plans.

5. 💡 Option 4 — Savings Plans

A modern alternative to RIs, committing to spend a fixed $ per hour over 1–3 years.

Features

Flexible across:
- Instance size
- OS
- Region
- Tenancy (default, dedicated, host)
Applies to EC2, Fargate, and Lambda usage.
Same savings as RIs (up to 72%).

✅ Recommended for most long-term, steady workloads.

6. 🏠 Option 5 — Dedicated Hosts

Get a physical EC2 server fully reserved for your account.

Use Cases

Compliance or regulatory isolation.
BYOL (Bring Your Own License) software (Oracle, SQL Server).
Control instance placement and underlying hardware.

Launch Steps

EC2 → Dedicated Hosts → Allocate Dedicated Host
Select:

Instance family (e.g., c5)
Availability Zone
1. Click Allocate

⚠️ Cost is much higher — typically for enterprise or compliance workloads.

7. 🧱 Option 6 — Capacity Reservations

Guarantee EC2 capacity in a specific Availability Zone (AZ) — even if you’re not running anything yet.

Features

Pay On-Demand price (no discount).
Reserve exact instance type & count.
Duration: open-ended or fixed end time.
Cancelling stops future billing, but you’re charged while reserved.

✅ Useful for:

Disaster recovery (DR)
Mission-critical systems
Short-term but guaranteed compute bursts

Example:

Reserve 4 × m5.2xlarge in eu-central-1a
→ You pay even if not used, but AWS guarantees capacity exists.

8. 🧠 Comparison Summary

Option	Pricing	Commitment	Flexibility	Use Case
On-Demand	High	None	Very high	Unpredictable workloads
Spot	Lowest	None	Moderate	Batch, analytics, non-critical
Reserved	Low	1–3 yrs	Fixed	Databases, web servers
Savings Plan	Low	1–3 yrs	High	Steady spend, variable workload
Dedicated Host	Very high	Optional	Low	Licensing, compliance
Capacity Reservation	On-Demand	None	Moderate	Guaranteed AZ capacity