AWS Batch & AWS Beanstalk

Beanstalk deployment models

-All at Once -Rolling -Rolling with Additional Batch -Immutable -Traffic Split

Types of Environment for AWS Beanstalk

Web Tier Environment and Worker Environment

Amazon Elastic Beanstalk

compute service for deploying and scaling applications (just write code, don't worry about infrastructure)

Worker Environment (Beanstalk)

Beanstalk installs a Daemon which pulls requests from SQS queue and executes tasks

What does AWS Batch execute on?

EC2 (including spot instances) and AWS Fargate

job definitions

define how job is supposed to be run (i.e. associated IAM role, vCPU requirement, container properties)

types of compute environments

managed (user gives min/max vCPU, etc.), and unmanaged (DIY)

job queues

reside in job queue where they wait till they are scheduled

What are the two categories of AWS Batch jobs?

simple or array (i.e. parametric sweeps/monte carlo)

Batch jobs tips

use Fargate and AWS batch generally, but use EC2 if work scale is large

Web Tier Environment (Beanstalk)

various EC2s in a ASG with SGs within an AZ connects to an ELB and the endpoint, also can connect to external DB