Public IPv4 Charges
New Charges Since Feb 1, 2024
AWS now charges for all Public IPv4 addresses whether used or not.
Pricing
| Metric | Cost |
|---|---|
| Per Hour | $0.005 |
| Per Month | ~$3.60 |
| Per Year | ~$43.80 |
Free Tier (EC2 Only)
| Account Created | Free Tier Instances | Duration |
|---|---|---|
| Before July 15, 2025 | t2.micro (or t3.micro where t2 unavailable) | 750 hours/month, first 12 months |
| On or after July 15, 2025 | t3.micro, t3.small, t4g.micro, t4g.small, c7i-flex.large, m7i-flex.large | 6 months or until credits used |
| Service | Free Tier for IPv4? |
|---|---|
| EC2 | ✅ See above |
| ELB | ❌ No free tier |
| RDS | ❌ No free tier |
| Other services | ❌ No free tier |
EC2 Free Tier Calculation:
750 hours = ~1 month of continuous running
├── 1 instance running 24/7 = 720 hours ✓
├── 2 instances running 24/7 = 1440 hours ✗ (over limit)
└── 3 instances running 8 hrs/day = 720 hours ✓
Services That Create Public IPs
Public IPv4 Sources:
├── EC2 Instances (with public IP enabled)
├── Elastic Load Balancers (1 IP per AZ)
├── NAT Gateways
├── RDS Instances (if publicly accessible)
├── Elastic IPs (even if not attached!)
└── Other services with public endpoints
Where public IPv4 charges come from
Monitoring IPv4 Usage
Method 1: Billing Dashboard
Billing → Bills → Select Month → Charges by Service
└── Look for "Public IPv4" charges
Method 2: IPAM (IP Address Manager)
VPC Console → IPAM → Public IP Insights
└── Create IPAM (free tier eligible)
└── View all public IPs across regions
Cost Optimization Tips
- Delete unused Elastic IPs - Charged even when not attached
- Use private subnets - Instances don't need public IPs
- Use NAT Gateway - Single public IP for multiple instances
- Consider IPv6 - No charges (but not all ISPs support it)
- Clean up after hands-on - Delete resources promptly
Spot Instances Deep Dive
How Spot Pricing Works
┌─────────────────────────────────────────────────────────┐
│ Spot Price Over Time │
│ │
│ $0.10 ─ ─ ─ ─ ─ On-Demand Price ─ ─ ─ ─ ─ ─ ─ ─ ─ │
│ │
│ $0.05 ════════ Your Max Price ════════════════════ │
│ │
│ $0.04 ┌──┐ ┌────┐ │
│ │ │ │ │ Spot Price │
│ $0.03 │ └────┘ └──┐ │
│ │ │ │
│ $0.02 └───────────────┴──────────────────────── │
│ │
│ ├─ Instance ─┤ ├─ Instance ─┤ │
│ Running Running │
│ │
│ When Spot Price > Max Price: 2-minute warning │
└─────────────────────────────────────────────────────────┘
Spot Price Characteristics
- Varies by instance type, AZ, and time
- Based on supply and demand
- Can check historical prices in console
- Different AZs have different prices
Interruption Handling
When spot price exceeds your max price:
| Option | Behavior | Use Case |
|---|---|---|
| Stop | Instance stops, EBS preserved | Resume work later |
| Terminate | Instance deleted | Stateless workloads |
| Hibernate | RAM saved to EBS, resume quickly | Quick restart needed |
Grace Period: 2 minutes to save work before interruption
Spot Request Types
One-Time Request
Request Created
│
▼
┌─────────────┐
│ Pending │
└──────┬──────┘
│ Capacity available
▼
┌─────────────┐
│ Active │──────▶ Instance Launched
└──────┬──────┘
│ Fulfilled
▼
┌─────────────┐
│ Closed │ Request complete, won't restart
└─────────────┘
Persistent Request
Request Created
│
▼
┌─────────────┐
│ Pending │◀──────────────────────┐
└──────┬──────┘ │
│ Capacity available │
▼ │
┌─────────────┐ │
│ Active │──────▶ Instance │
└──────┬──────┘ Launched │
│ │ │
│ │ Interrupted
│ │ │
│ ▼ │
│ Instance │
│ Stopped/ │
│ Terminated │
│ │ │
└────────────────────┴─────────┘
Auto-restart when price allows
Spot request lifecycle
Terminating Spot Instances Correctly
Critical Order
Wrong order = instances keep relaunching!
Correct Order:
- Cancel the Spot Request first
- Then terminate the instances
Wrong Way:
Terminate Instance → Request sees 0 instances → Launches new instance!
Right Way:
Cancel Request → Request disabled → Terminate Instance → Done!
Spot Request States
| State | Description |
|---|---|
open |
Request submitted, waiting for capacity |
active |
Request fulfilled, instances running |
disabled |
Request stopped (persistent only) |
cancelled |
Request cancelled by user |
closed |
Request completed (one-time) |
failed |
Request failed (bad parameters) |
Can Cancel When: open, active, or disabled
Spot Fleets
What is a Spot Fleet?
Collection of Spot Instances (+ optional On-Demand) that automatically maintains target capacity.
Spot Fleet
├── Launch Pool 1: c5.large, us-east-1a
├── Launch Pool 2: c5.large, us-east-1b
├── Launch Pool 3: m5.large, us-east-1a
├── Launch Pool 4: m5.large, us-east-1b
└── Target: 10 instances (or vCPUs, or memory)
Allocation Strategies
| Strategy | How It Works | Best For |
|---|---|---|
| lowestPrice | Launches from cheapest pool | Maximum cost savings |
| diversified | Spreads across all pools | High availability |
| capacityOptimized | Launches from pool with most capacity | Reducing interruptions |
| priceCapacityOptimized | Best capacity first, then lowest price | Recommended for most |
Spot Fleet Configuration
Spot Fleet Request:
├── Launch Template / Specifications
│ ├── AMI
│ ├── Instance Types (multiple)
│ └── Subnets (multiple AZs)
├── Target Capacity
│ ├── Type: instances / vCPUs / memory
│ └── Amount: 10
├── Allocation Strategy: priceCapacityOptimized
├── On-Demand Base Capacity: 2 (optional)
└── Maintain Target Capacity: Yes
Spot Fleet vs Spot Request
| Feature | Spot Request | Spot Fleet |
|---|---|---|
| Instance Types | Single | Multiple |
| AZs | Single | Multiple |
| Allocation Strategy | N/A | Yes |
| Auto-replace interrupted | Persistent only | Yes |
| Mix with On-Demand | No | Yes |
Hands-On: Spot Request UI
Viewing Spot Pricing History
EC2 Console → Spot Requests → Pricing History
├── Select Instance Type: c4.large
├── Select OS: Linux/Unix
├── Select Time Range: 3 months
└── View price trends by AZ
Creating Spot Request
EC2 Console → Instances → Launch Instance
└── Advanced Details
└── Request Spot Instances ✓
├── Max Price: (default = On-Demand price)
├── Request Type: One-time / Persistent
├── Valid Until: (for persistent)
└── Interruption Behavior: Terminate / Stop / Hibernate
Questions & Answers
Q1: How much does a Public IPv4 cost per month?
Answer:
~$3.60/month ($0.005/hour × 24 hours × 30 days).
EC2 has 750 hours/month free tier (first 12 months). Other services like ELB and RDS have no free tier for IPv4.
Q2: What happens when Spot price exceeds your max price?
Answer:
You get a 2-minute warning, then your instance is either:
- Stopped (EBS preserved, can restart later)
- Terminated (instance deleted)
- Hibernated (RAM saved to EBS)
You choose the behavior when creating the request.
Q3: What's the difference between Spot Request and Spot Fleet?
Answer:
- Spot Request: Single instance type, single AZ
- Spot Fleet: Multiple instance types, multiple AZs, allocation strategies, can mix with On-Demand
Spot Fleet is more flexible and can automatically choose the best pools.
Q4: Which Spot Fleet allocation strategy is recommended?
Answer:
priceCapacityOptimized - It first selects pools with highest capacity (reducing interruption risk), then chooses the lowest price among those. Best balance of cost and availability.
Q5: Why must you cancel Spot Request before terminating instances?
Answer:
For persistent Spot Requests, the request monitors instance count. If you terminate instances first, the request sees "0 instances" and automatically launches new ones to meet target capacity.
Cancel request first → Request disabled → Then terminate instances safely.
Check yourself
Quiz
Flashcards
Click the card (or focus it), then use ← → to move and Space to flip.