cortexlabs / cortex

Production infrastructure for machine learning at scale

Home Page:https://cortexlabs.com/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Use EC2 Warm Pools to reduce the scale-out time

RobertLucian opened this issue · comments

Description

Helps reduce the scale-out time.

Additional

https://docs.aws.amazon.com/autoscaling/ec2/userguide/ec2-auto-scaling-warm-pools.html
https://medium.com/keikoproj/rapid-auto-scaling-on-eks-part-2-d094b9b2cd62

Notes

On the pricing model:

You have the option of keeping instances in the warm pool in one of two states: Stopped or Running. Keeping instances in a Stopped state is an effective way to minimize costs. With stopped instances, you pay only for the volumes that you use and the Elastic IP addresses that are not assigned to a running instance. But you don't pay for the stopped instances themselves. You pay for the instances only when they are running.