A Python script to create an EC2 Spot instance using boto3

How do you create EC2 Spot Instance

A Python script to create an EC2 Spot instance using boto3. The script launches a spot instance based on an AMI, associates it with a specific security group in a given VPC, and assigns custom tags for easy identification.

Prerequisites:

• Install boto3: pip install boto3
• Configure AWS CLI credentials: aws configure

import boto3
import time

# AWS Configuration
region = "us-east-1"  # Change as needed
ami_id = "ami-xxxxxxxxxxxxxxxxx"  # Replace with your AMI ID
instance_type = "t3.micro"  # Change instance type as needed
security_group_id = "sg-xxxxxxxxxxxxxx"  # Replace with your security group ID
subnet_id = "subnet-xxxxxxxxxxxxxx"  # Replace with your subnet ID within the VPC
key_name = "your-key-pair"  # Replace with your SSH key pair name
tag_name = "MySpotInstance"

# Initialize boto3 clients
ec2_client = boto3.client("ec2", region_name=region)

# Request Spot Instance
spot_request = ec2_client.request_spot_instances(
    InstanceCount=1,
    LaunchSpecification={
        "ImageId": ami_id,
        "InstanceType": instance_type,
        "KeyName": key_name,
        "SecurityGroupIds": [security_group_id],
        "SubnetId": subnet_id,
    },
    Type="one-time",
    SpotPrice="0.005",  # Adjust price based on instance type and AWS pricing
)

# Extract Spot Request ID
spot_request_id = spot_request["SpotInstanceRequests"][0]["SpotInstanceRequestId"]
print(f"Spot Instance Request ID: {spot_request_id}")

# Wait for Spot Request to be fulfilled
while True:
    response = ec2_client.describe_spot_instance_requests(SpotInstanceRequestIds=[spot_request_id])
    state = response["SpotInstanceRequests"][0]["State"]
    print(f"Spot request state: {state}")

    if state == "active":
        instance_id = response["SpotInstanceRequests"][0]["InstanceId"]
        print(f"Spot Instance Launched: {instance_id}")
        break
    elif state in ["cancelled", "failed", "closed"]:
        print("Spot request failed or was cancelled.")
        exit(1)

    time.sleep(5)  # Wait and check again

# Assign Tags to the Instance
ec2_client.create_tags(
    Resources=[instance_id],
    Tags=[
        {"Key": "Name", "Value": tag_name},
        {"Key": "Environment", "Value": "Test"},
        {"Key": "Owner", "Value": "YourName"},
    ],
)

print(f"Instance {instance_id} tagged successfully!")

Explanation:

1. Request a Spot Instance: Uses request_spot_instances() to request a spot instance with a specified AMI, instance type, security group, and subnet.
2. Monitor the request: The script polls AWS every 5 seconds until the instance is assigned.
3. Assign tags: Tags are applied for easier searching and organization.

Search EC2 Instances with Tag Information

You can search for EC2 instances by tags using boto3 in Python. The describe_instances method allows filtering instances based on specific tags.

Example: Search EC2 Instances by Tag

This script retrieves EC2 instances with a specific tag and prints their instance IDs and details.

Python Script

import boto3

# AWS Configuration
region = "us-east-1"  # Change as needed
tag_key = "Name"  # Change to the tag key you are searching for
tag_value = "MySpotInstance"  # Change to the tag value you are searching for

# Initialize EC2 client
ec2_client = boto3.client("ec2", region_name=region)

# Search EC2 instances based on a tag
response = ec2_client.describe_instances(
    Filters=[
        {"Name": f"tag:{tag_key}", "Values": [tag_value]},
        {"Name": "instance-state-name", "Values": ["running", "pending"]},  # Optional: Filter by state
    ]
)

# Extract instance details
instances = []
for reservation in response["Reservations"]:
    for instance in reservation["Instances"]:
        instances.append({
            "Instance ID": instance["InstanceId"],
            "State": instance["State"]["Name"],
            "Type": instance["InstanceType"],
            "Public IP": instance.get("PublicIpAddress", "N/A"),
            "Private IP": instance["PrivateIpAddress"],
        })

# Display results
if instances:
    for instance in instances:
        print(instance)
else:
    print("No instances found with the given tag.")

How It Works:

1. Uses describe_instances() with a filter on tag:Name = MySpotInstance.
2. Filters only running or pending instances (you can modify this).
3. Extracts instance details (ID, state, type, IP addresses).
4. Displays results in a structured format.

Example Output:

{'Instance ID': 'i-0123456789abcdef0', 'State': 'running', 'Type': 't3.micro', 'Public IP': '34.123.45.67', 'Private IP': '10.0.1.12'}

Terminate EC2 Instances

You can terminate an EC2 instance using boto3 in Python by calling the terminate_instances() method.

Python Script to Terminate an EC2 Instance

import boto3

# AWS Configuration
region = "us-east-1"  # Change as needed
instance_id = "i-0123456789abcdef0"  # Replace with your EC2 Instance ID

# Initialize EC2 client
ec2_client = boto3.client("ec2", region_name=region)

# Terminate the instance
response = ec2_client.terminate_instances(InstanceIds=[instance_id])

# Extract termination details
for instance in response["TerminatingInstances"]:
    print(f"Instance {instance['InstanceId']} is {instance['CurrentState']['Name']}")

How It Works

1. Uses terminate_instances() to terminate the specified EC2 instance.
2. Prints the instance ID and its current state.

Example Output

Instance i-0123456789abcdef0 is shutting-down

Terminate EC2 Instances by Tag

If you want to terminate instances based on a tag, use this script:

# Terminate EC2 instances by tag
tag_key = "Name"
tag_value = "MySpotInstance"

# Search for instances with the given tag
instances = ec2_client.describe_instances(
    Filters=[{"Name": f"tag:{tag_key}", "Values": [tag_value]}]
)

instance_ids = [
    instance["InstanceId"]
    for reservation in instances["Reservations"]
    for instance in reservation["Instances"]
]

if instance_ids:
    response = ec2_client.terminate_instances(InstanceIds=instance_ids)
    for instance in response["TerminatingInstances"]:
        print(f"Instance {instance['InstanceId']} is {instance['CurrentState']['Name']}")
else:
    print("No instances found with the given tag.")

Precautions

• Cannot undo termination: Once an EC2 instance is terminated, you cannot recover it.
• Detach important volumes: If your instance has valuable EBS volumes, detach them first.
• Verify instance ID: Double-check that you're terminating the correct instance.