AWS-Logo_White-Color
1
Launching EC2 Spot Instances
Starting the workshop
Creating a Launch Template
Launching EC2 Spot Instances via EC2 Auto Scaling group
Launching EC2 Spot Instances via EC2 Fleet
Creating a Spot Interruption Experiment
Savings Summary
Spot Placement Score (SPS)
Spot Blueprints
(Optional) - Launching an EC2 Spot Instance via the RunInstances API
(Optional) - Launching an EC2 Spot Instance via Spot Fleet request
Learn more
Clean Up
2.
EC2 Auto Scaling
Requirements and Notes
Architecture
Starting the workshop
...At an AWS event
...On your own
Setup with CloudFormation
Using the Cloud9 Environment
Setup AWS CLI and other tools
Clone the GitHub repo
Create an EC2 launch template
Deploy the AWS Elastic Load Balancer
Create an EC2 Auto Scaling Group
Browse to the web app
Using Dynamic Scaling in ASG
Stress the app with AWS Systems Manager
Resilience with Spot Instances
Finished!
Cleanup
Learn more
3.
Efficient EC2 Auto Scaling
Starting the workshop
...At an AWS event
...On your own
Amazon EC2 Auto Scaling
Predictive scaling
Working with Custom Metrics
Create Predictive Scaling Policy
Review the outcome of predictive scaling
Configure Dynamic scaling
Take control of instance Lifecycle
Measure the Launch Speed of Instances
Warm pools
How Warm Pools work?
Launch From Warm Pool
Instance Refresh and Rollback
Cleanup
Finished!
Learn more
4.
EKS with Karpenter
Start the workshop
...on your own
...at an AWS event
Test the Cluster
Explore Kubernetes Tools
Explore Kube-ops-view
Karpenter
Explore Karpenter Installation
Set up the Provisioner
Automatic Node Provisioning
Consolidation
Deploying Multiple Provisioners
EC2 Spot deployments
Multi-Architecture deployments
Using Alternative Provisioners
Scaling App and Cluster
Build the Microservice
Deploy the Microservice
Configure HPA
Stress-test the sytem
Use FIS to Interrupt a Spot Instance
Cleanup
Conclusion
What Have We Accomplished
5.
EKS with Cluster Autoscaler
Start the workshop
...at an AWS event
...on your own
Test the Cluster
Install Kube-ops-view
Add EKS managed Spot workers
EKS managed node groups
Select Instance Types for Diversification
Create EKS managed node groups with Spot capacity
Spot Best Practices and Interruption Handling
Deploy an example Microservice
Monte Carlo Pi Template
Tolerations and Affinity
Deploy Application
Autoscale Cluster and Application
Configure Cluster Autoscaler (CA)
Scale Cluster with CA
Configure Horizontal Pod Autoscaler (HPA)
Stress test the Application
(Optional) Add self managed Spot workers
Create self managed node groups with Spot Instances
Deploy AWS Node Termination Handler
Test Autoscaling of Cluster and Application
Self managed Spot workers cleanup
(Optional) Running cost optimized and resilient Jenkins jobs
Create Spot workers for Jenkins
Setting up Jenkins server
Increasing resilience
Running Jenkins jobs
Auto scaling Jenkins nodes
Jenkins cleanup
Cleanup
Conclusion
What Have We Accomplished
6.
DoEKS: Spark apps with EKS
7.
Containers with ECS
Prerequisites
Introduction
Introduction to Containers
Introduction to ECS
Scaling ECS Workloads
Starting the workshop
...At an AWS event
...On your own (self-paced)
Setup the Workspace environment
Deploy CloudFormation Stack
Setup Cloud9 Environment
Resize Cloud9 Instance Root Volume
Setup AWS CLI and clone the workshop repo
Using Spot Instances with Auto Scaling groups capacity providers
Architecture
Create an ECS cluster
Create On-Demand Auto Scaling Group and Capacity Provider
Selecting Spot Instance Types
Create EC2 Spot Auto Scaling Group and Capacity Provider
Setup default capacity provider strategy
Visualizing ECS Scaling Metrics
Create an ECS service
ECS Managed Scaling
EC2 Spot Interruption Handling in ECS
Using AWS Fargate Spot capacity providers
Add Fargate capacity providers to ECS Cluster
Create ECS Fargate Tasks
Create ECS Fargate service
Cleanup
Conclusion
8.
Running Spark apps with EMR on Spot Instances
Starting the workshop
...at an AWS event
...on your own
EMR Instance Fleets
Right sizing Spark executors
Selecting instance types
Launch a cluster - Step 1
Launch a cluster - Step 2
Launch a cluster - Steps 3&4
Fleet configuration options
Examining the cluster
Tracking Spot interruptions
Creating the Spot Interruption Experiment
Interrupting a Spot Instance
Verifying the app's results
Analyzing costs
Cleanup
Call to Action
9.
Pricing Financial Derivatives with AWS Batch
Architecture
Starting the workshop
... At an AWS event
... On your own
Risk pipeline
Creating your Docker image
AWS Batch
Creating the compute environment
Creating the job queue
Registering the job definition
Workflow orchestration
Monitoring and results
Clean Up
Final Thoughts
10.
Rendering with AWS Batch
Architecture
Starting the workshop
... At an AWS event
... On your own
Rendering pipeline
Creating your Docker image
AWS Batch
Creating the compute environment
Creating the job queue
Registering the job definition
Workflow orchestration
Monitoring and results
AWS Fault Injection Simulator
Creating the experiment
Starting a new rendering job
Running the experiment
Monitoring the experiment
Clean Up
Final Thoughts
11.
CI/CD and Test Workloads
GitLab
Starting the workshop
...At an AWS event
...On your own
Workshop Preparation
Create a GitLab repository
Configure GitLab runners on Spot instances
Building the demo app
Deploying Amazon EKS on Spot instances
Installing the demo app into Amazon EKS
Workshop Cleanup
12.
EC2 Workloads at Scale
Requirements and Notes
Architecture
Starting the workshop
...At an AWS event
...On your own
Setup with CloudFormation
Using the Cloud9 Environment
Setup AWS CLI and other tools
Clone the GitHub repo
Create an EC2 launch template
Deploy the database with Amazon RDS
Deploy the AWS Elastic Load Balancer
Create an EC2 Auto Scaling Group
Seed the database with application data
Deploy to the ASG with CodeDeploy
Using Dynamic Scaling in ASG
Stress the app with AWS Systems Manager
Increasing resilience when using Spot Instances
Finished!
Cleanup
Learn more
More
Amazon EC2 Spot Instances
Authors
Resources
Feedback / Questions?
English
Clear History
Github Repo
EC2 Spot Workshops
Star
Fork
Site Info
Privacy
Site Terms
© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Edit this page
Amazon EC2 Spot Workshops
>
General guidance
- [Blog post] New Amazon EC2 Spot pricing model: Simplified purchasing without bidding and fewer interruptions
- [Code] EC2 Spot Interruption Handler for Auto Scaling groups and Spot Fleet
- [Code] Amazon EC2 Metadata Mock (Spot Interruption Simulation testing)
- [Blog post] New – EC2 Auto Scaling Groups With Multiple Instance Types & Purchase Options
- [Blog post] Capacity-Optimized Spot Instance Allocation in Action at Mobileye and Skyscanner
- [Blog post]: Proactively manage the Spot Instance lifecycle using the new Capacity Rebalancing feature for EC2 Auto Scaling
- [Blog post]: Attribute-Based Instance Type Selection for EC2 Auto Scaling and EC2 Fleet
- [Blog post]: Identifying optimal locations for flexible workloads with Spot placement score
- [Blog post]: Implementing interruption tolerance in Amazon EC2 Spot with AWS Fault Injection Simulator
- [AWS Online Tech Talks]: Optimize AI/ML, Kubernetes and Big Data Workloads at Your Startup
Web applications running on EC2
- [Blog post] Running Web Applications on Amazon EC2 Spot Instances
- [Tutorial] Amazon EC2 Auto Scaling with EC2 Spot Instances
- [AWS Online Tech Talks]: How to Build Scalable Web Based Applications for Less with Amazon EC2 Spot Instances
Big-data and analytics
- [Blog post] Optimizing Amazon EMR for resilience and cost with capacity-optimized Spot Instances
- [Blog post] Best practices for running Apache Spark applications using Amazon EC2 Spot Instances with Amazon EMR
- [Tutorial] Getting started tutorial: Optimizing Amazon EMR clusters for cost and scale with EC2 Spot Instances and Amazon EMR
- [AWS re:Invent 2019 Talk] Run big data workloads faster and cheaper (20 minute demo - EMR Instance Fleets)
- [AWS Online Tech Talks]: Best Practices for Running Spark Applications Using Spot Instances on EMR
- [Blog post] Running cost optimized Spark workloads on Kubernetes using EC2 Spot Instances
Containers
- [Blog post] Amazon EKS now supports provisioning and managing EC2 Spot Instances in managed node groups
- [Blog post] Building for Cost optimization and Resilience for EKS with Spot Instances
- [Tutorial] Getting started tutorial: Run Kubernetes clusters for less with Amazon EKS and Spot Instances
- [AWS Online Tech Talks] Save Costs Running Kubernetes Clusters with EC2 Spot Instances
- [Blog post] Deep dive into Fargate Spot to run your ECS Tasks for up to 70% less
- [Code] AWS Node Termination Handler for Kubernetes
- [Blog post]: Efficiently Scaling kOps clusters with Amazon EC2 Spot Instances
- [Tutorial]: Run kOps Kubernetes clusters for less with Amazon EC2 Spot Instances
AI/ML
- [Tutorial] Getting started tutorial: Optimizing and Scaling Machine Learning Training with Managed Spot Training for Amazon SageMaker
- [Blog post] Train Deep Learning Models on GPUs using Amazon EC2 Spot Instances
- [AWS On-line Tech Talks] Reduce Inferencing Cost by Up to 90% Using Amazon Elastic Inference and Amazon EC2 Spot Instances
HPC & Batch workloads
- [Blog post] Western Digital HDD Simulation at Cloud Scale – 2.5 Million HPC Tasks, 40K EC2 Spot Instances
- [Tutorial] Running batch jobs at scale for less with AWS Batch and EC2 Spot Instances
- [AWS Online Tech Talks] Optimize Batch Processing for Scale & Save Up to 90% on Compute Running AWS Batch on Spot Instances
CI/CD
- [Blog post] Cost Optimize your Jenkins CI/CD pipelines using EC2 Spot Instances
- [Tutorial] Cost optimize Jenkins jobs on Kubernetes with EC2 Spot Instances
Site Info
Privacy
Site Terms