AWS at Slack
     at Slack

By Richard Crowley

WhoTF am I?

Substrate

https://substrate.tools/

The right way to use AWS, designed for startups with immediate security and reliability needs, and informed by everything I did at Slack

AWS at Tiny Speck in 2009

AWS at Tiny Speck Slack in 2014

“We only use EC2 and S3, so we can leave whenever we want”

(No one in particular said that but the sentiment was in the air.)

EC2 is lock-in

EC2 is lock-in

The peculiarities of the networks make EC2 and GCE fundamentally different

Rate limits and service quotas

Rate limits and service quotas

The best way to avoid running into rate limits and service quotas is to have lots of AWS accounts

Reserved Instances bin packing

Reserved Instances bin packing

Save yourself less money but a ton of time by buying Savings Plans instead of Reserved Instances

Backups and existential dread

Backups and existential dread

Store your backups in a separate AWS account and don’t give anyone enough privileges to delete them

SecOps

SecOps

Use an AWS account boundary to ensure that your all-seeing security eyes can’t themselves be seen

Futile separation of dev and prod

Futile separation of dev and prod

Accounts are the one true unit of isolation in AWS

2015 security incident and the move to VPC

2015 security incident and the move to VPC

Security monitoring, security audits, and disaster recovery exercises are very important

Kubernetes vs Mesos vs me sitting on everyone’s hands

Kubernetes vs Mesos vs me sitting on everyone’s hands

Not only did waiting ensure we chose the winning path but we reaped the benefits of tooling advances like EKS that the earliest adopters didn’t

Project White Castle

Project White Castle

Accounts are the one true unit of isolation in AWS and VPC Sharing is the smartest way to network them back together

AWS at Slack in 2020

Takeaways

Thank you!

Questions?

rcrowley@rcrowley.org