Amazon EC2 Best Instance (amazon-ec2-best-instance)
Amazon EC2 Best Instance (amazon-ec2-best-instance) allows you to choose the most optimal and cheap EC2 instance type for on-demand and spot and with a less reclaimed rate for a spot instance.
Deprecation Notice
Important: This package is deprecated and should no longer be used. Please switch to the 'best-ec2' package by running:
pip install best-ec2
https://pypi.org/project/best-ec2/
Prerequisites
- python3
- pip3
- boto3
- AWS Account
- AWS Credentials
Install
pip install amazon-ec2-best-instance
Options
- vcpu Required. Float. Describes the vCPU configurations for the instance type.
- memory_gb Required. Float. Describes the memory for the instance type in GiB.
- usage_class Optional. String. Indicates whether the instance type is offered for spot or On-Demand. Default: 'on-demand'. Values: 'spot'|'on-demand'.
- burstable Optional. Boolean. Indicates whether the instance type is a burstable performance instance type.
- architecture Optional. String. The architectures supported by the instance type. Default: 'x86_64'. Values: 'i386'|'x86_64'|'arm64'|'x86_64_mac'
- product_descriptions Optional. List. The operating system that you will use on the virtual machine. Default: ['Linux/UNIX']. Values: Linux/UNIX | Red Hat Enterprise Linux | SUSE Linux | Windows | Linux/UNIX (Amazon VPC) | Red Hat Enterprise Linux (Amazon VPC) | SUSE Linux (Amazon VPC) | Windows (Amazon VPC)
- is_current_generation Optional. Boolean. Use the latest generation or not.
- has_gpu Optional. Boolean. Indicating whether the instance is equipped with a GPU. Default: None
- gpu_memory Optional. Integer. Amount of GPU Memory in Gigabytes (GiB). Only activated when 'has_gpu' is set to True. Default: None
- is_instance_storage_supported Optional. Boolean. Use instance types with instance store support
- max_interruption_frequency Optional. Integer (%). Max spot instance frequency interruption in percent. Note: If you specify >=21, then the '>20%' rate is applied. It is used only if 'usage_class' == 'spot' and 'is_best_price' == True
- availability_zones Optional. List. Availability zones. E.g. ['us-east-1a', 'us-east-1b']
- final_spot_price_determination_strategy Optional. String. Default: "min". Valid values: "min"|"max"|"average"
Usage
Simple
from amazon_ec2_best_instance import Ec2BestInstance
ec2_best_instance = Ec2BestInstance()
# It returns all available instance types, including those with over-provisioning resources (CPU, memory, etc.).
response = ec2_best_instance.get_best_instance_types({
'vcpu': 1,
'memory_gb': 2
})
print(response)
'''
[{'instance_type': 'c5n.2xlarge'}, ... , {'instance_type': 'x2iedn.8xlarge'}]
'''
Advanced
import boto3
from botocore.config import Config
import logging
from amazon_ec2_best_instance import Ec2BestInstance
ec2_client_config = Config(
retries={
'max_attempts': 20,
'mode': 'adaptive'
}
)
pricing_client_config = Config(
retries={
'max_attempts': 10,
'mode': 'standard'
}
)
ec2_client = boto3.Session().client('ec2', config=ec2_client_config)
pricing_client = boto3.Session().client('pricing', config=pricing_client_config)
# Optional.
options = {
# Optional. Default: us-east-1
'region': 'us-east-1',
# Optional. Default: 10
'describe_spot_price_history_concurrency': 20,
# Optional. Default: 10
'describe_on_demand_price_concurrency': 20,
'clients': {
'ec2': ec2_client,
'pricing': pricing_client
},
# Optional. Integer. Default: 120. It limits the lifetime of cache data.
'cache_ttl_in_minutes': 60
}
logging.basicConfig(level=logging.INFO, format='%(asctime)s: %(levelname)s: %(message)s')
# Optional.
logger = logging.getLogger()
ec2_best_instance = Ec2BestInstance(options, logger)
response = ec2_best_instance.get_best_instance_types({
# Required.
'vcpu': 1,
# Required.
'memory_gb': 2,
# Optional. Default: 'on-demand'. Values: 'spot'|'on-demand'
'usage_class': 'spot',
# Optional.
'burstable': False,
# Optional. Default: 'x86_64'. Values: 'i386'|'x86_64'|'arm64'|'x86_64_mac'
'architecture': 'x86_64',
# Optional. Default: ['Linux/UNIX'].
# Values: Linux/UNIX | Red Hat Enterprise Linux | SUSE Linux | Windows | Linux/UNIX (Amazon VPC) |
# Red Hat Enterprise Linux (Amazon VPC) | SUSE Linux (Amazon VPC) | Windows (Amazon VPC)
'product_descriptions': ['Linux/UNIX'],
# Optional.
'is_current_generation': True,
# Optional. If this parameter is set to True, the method will return the instance type with the best price.
'is_best_price': True,
# Optional. If this parameter is set to True, the method will return the instance type with the instance storage.
'is_instance_storage_supported': True,
# Optional. Integer. Max spot instance frequency interruption in percent.
'max_interruption_frequency': 10,
# Optional. List<String>. The availability zones.
'availability_zones': ['us-east-1a', 'us-east-1b']
})
print(response)
'''
[{'instance_type': 'c5d.large', 'price': '0.032700', 'interruption_frequency': {'min': 0, 'max': 5, 'rate': '<5%'}}, ...]
'''
Spot
If you need to get a spot instance with minimal price and minimal frequency of interruption you can use 'is_best_price' and/or 'max_interruption_frequency' input parameter
from amazon_ec2_best_instance import Ec2BestInstance
ec2_best_instance = Ec2BestInstance()
response = ec2_best_instance.get_best_instance_types({
# Required. Float
'vcpu': 31.2,
# Required. Float
'memory_gb': 100.5,
# Optional. String. Default: 'on-demand'. Values: 'spot'|'on-demand'
'usage_class': 'spot',
# Optional. Boolean.
# If this parameter is set to True, the method will return the instance type with the best price.
'is_best_price': True,
# Optional. Integer. Max spot instance frequency interruption in percent.
# Note: If you specify >=21, then the '>20%' rate is applied
# It is used only if 'usage_class' == 'spot' and 'is_best_price' == True
'max_interruption_frequency': 10
})
print(response)
'''
[{'instance_type': 'm6id.8xlarge', 'price': '0.642600', 'interruption_frequency': {'min': 6, 'max': 10, 'rate': '5-10%'}}, ...]
'''
from amazon_ec2_best_instance import Ec2BestInstance
ec2_best_instance = Ec2BestInstance()
response = ec2_best_instance.get_best_instance_types({
# Required. Float
'vcpu': 31.2,
# Required. Float
'memory_gb': 100.5,
# Optional. String. Default: 'on-demand'. Values: 'spot'|'on-demand'
'usage_class': 'spot',
# Optional. Boolean.
# If this parameter is set to True, the method will return the instance type with the best price.
'is_best_price': True
})
print(response)
'''
[{'instance_type': 'r5a.8xlarge', 'price': '0.578100'}, ...]
'''