What are EC2 Spot Instances
A Spot Instance utilizes spare EC2 capacity, often priced well below the On-Demand rate. Leveraging these instances allows for substantial cost reduction in your Amazon EC2 usage. They are an economical option, especially when your application’s runtime is flexible. Spot Instances excel in scenarios like data analysis, batch processing, background tasks, and non-urgent workloads.
How Spot Instances Work
Infrastructure
Spare EC2 Capacity
Provisioned from spare capacity, uses same infrastructure as On Demand (OD).
Pricing
Up to 90% off compared to On Demand
Pricing is based on long term supply and demand, smooth and predictable.
Interruption
Only interrupted if OD needs capacity
AWS can reclaim with 2 minute notice by issuing two types of notifications to help handle interruptions.
Diversification
and flexibility is key
Make use of different instance types, sizes, Availability Zones, and times.
Amazon EC2 Pricing Options
On-Demand
Pay for compute capacity by the second with no long-term commitments.
Â
Spiky stateful workloads,
to define needs
Reserved Instances / Savings Plan
Make a 1 or 3 year commitment and receive a significant discount off On-Demand prices.
Â
Committed and
steady-state usage
EC2 Spot
Spare Amazon EC2 capacity at savings of up to 90% off On-Demand prices.
Â
Fault-tolerant, flexible, stateless workloads
Benefits of EC2 Spot
Price Protection
Lock in cost savings with price protection mechanisms.
Instance Flexibility
Easily switch between instance types as needed.
Mixed Instance Groups
Combine different instance types for optimal performance and cost-efficiency.
Future Proof
Stay prepared for future workloads with adaptable Spot Instances.
AWS Management
Console, CLI, and SDKs: Manage Spot Instances seamlessly with AWS tools.
Preview Instances
Explore new instance types and features with preview instances.
Interruptions Handling
By the nature of Spot as spare-capacity, a Spot instance can be interrupted if the instance is needed by On-Demand. AWS provides two types of notifications to handle the response in an automated way:
EC2 instance rebalance recommendation
â–ªAct when your Spot instance is at elevated risk of interruption.
â–ª Built in support for integrations such as EC2 Auto
Scaling and EKS Managed Node Groups.
Spot instance termination notice (reactive)
â–ªInterruption of instance will
happen in 2 minutes, adjust
your workload appropriately
â–ª Built in support for AWS
integrations such as EC2 Auto
Scaling and EKS
Managed Node Groups +
What is an Ideal Spot Workload?
Ideal Spot workloads are those with flexible scheduling needs, allowing them to harness cost-effective EC2 capacity without strict time constraints. Workloads suitable for Spot Instances exhibit cost sensitivity and can gracefully handle interruptions.
- Fault-tolerant
- Flexible
- Loosely coupled
- Stateless
Web services
Containers
High Performance Compute + Batch
CI/CD
Big Data
AI/ML
Want to leverage Amazon EC2 Spot Instances?
Take your company to the next level with our DevOps and Cloud solutions
Availability: Monday – Friday 9:30 AM To 7:30 PM
Sales Enquiry
Career Opportunities
Support Enquiries
Frequently asked questions
Yes Definitely there are some tools that can help you in adoption of Spot instances:
- Spot Blueprints Dynamic infrastructure code generator, providing best-practice reference architecture on EC2 Spot usage with AWS services
- Attribute Based Instance Selection Rather than selecting specific instance types, this tool allows you to provision capacity on instance attributes (such as CPU/Memory) depending on requirements, also supporting new instances as soon as they become available.
- Find the optimal location (Reg/AZ) for your workload. You provide the resource & capacity requirements and SPS recommends locations (Reg/AZ) and the likelihood of getting the capacity you need.
- Enables chaos engineering and supports imitation of Spot interruptions, enabling you to prove the resilience of workloads running on Spot.
Savings can be substantial, often up to 90% compared to On-Demand pricing. However, the exact savings depend on factors like instance type, region, and current demand.
You can maintain availability by using Spot Fleet, a feature that automatically distributes your workload across multiple Spot Instances, Availability Zones, and instance types to minimize the risk of interruption.
Spot Instances can be interrupted if the capacity is needed elsewhere. To handle this gracefully, you can use tools like Amazon EC2 Auto Scaling to replace interrupted instances or save work progress before termination, ensuring minimal disruption to your application.
Copyright 2022