Chevre, this week’s parsha is Bo (בא). This week’s parsha starts off with the final 3 plagues. Locusts, Darkness and death of the firstborn.
The plague of Locusts is that they come and eat everything. In the cloud world we would call this a customer or user. They will come and use up everything offered to them. We also might call them a noisy neighbor. Someone else that has an instance or tenant on the same physical droplet (machine) that you are using and consuming more than their fair share of the resources. To be fair, they and you are entitled to those resources (most of the time). And the cloud provider has guard rails in place to ensure that “innocent” users are not impacted by a noisy neighbor. But it’s something that engineering needs to account for.
Darkness was interesting. The Israelites had light while the Egyptians had darkness. For 3 days!! And it wasn’t just difficult to see, the Egyptians were essentially blind for 3 days. In our world, that is what happens when our observability systems go down and we stop recording logs. We try to have systems in place to handle every possible failure scenario, but it’s a difficult engineering challenge. I know my personal home-lab systems have interesting failure modes based on where things are running (on-prem/at-home vs on EC2). And when some things are down and others are not, especially the authentication system, that can make things extra special challenging.
The final plague is Death of the Firstborn. For Paro, for a while at least, this is what forced him to allow the Israelites to leave Egypt. In the cloud world, things die all the time. It’s the difference between cattle and pets. Pets are when you manually fine-tune a system individually and it’s critical to you. Cattle is when you have hundreds, thousands or even millions of machines and 1 failing is barely a blip on the radar. At that scale failures are expected and you have common mechanisms and automations to handle the failures and work around them. Mai-Lan Tomer Bukovec spoke about the scale, engineering and regular hardware failures at S3 in this interview.
This interview re-ignited my appreciation for the scale, complexity and awesomeness of the compute cloud that we use every day.
Shabbat Shalom
https://www.youtube.com/watch?v=5vL6aCvgQXU&&pp=ygUWbWFpLWxhbiB0b21zZW4gYnVrb3ZlYw%3D%3D