Chapter 13 - High Availability and Scaling

Question 1

highly available?

Answer 1

Gear ourselves towards picking answers that are highly available,
unless you are presented with a situation that explicitly calls out that you do not need high availability and it is only focusing on cost, we always want to make sure that our resources can withstand failure
and are spread out across those multiple availability zones.

Question 2

horizontal scaling

Answer 2

We generally want to focus on horizontal scaling

• We can continue to spin up more resources
• More architecture
• It helps us be more highly available

Where vertical scaling just resizes that instance, has that natural limit.

Question 3

vertically scale

Answer 3

You might see a situation which says, we want to increase that network throughput for that t2.micro instance.

• The only real answer there is you need to scale vertically.
• You need to up size that to be something that’s a bit more appropriate.

Question 4

Auto Scaling

Answer 4

Is only for EC2 instances

• Auto Scaling will automatically recover that architecture.
• So no other service can be added into an Auto Scaling group.
• It cannot be used to add in load balancers
• or RDS databases
• or Lambda functions
• or anything else.
• Has that min, maxed, and desired capacity of one.
• This is great for that situation where maybe we have that legacy instance
• that NAT instance
• that resource where there can’t be more than one at a time.

Question 5

AMI (Amazon Machine Images)

Answer 5

• Loading it up with our dependencies and our code,
• They’re going to provision a lot quicker than trying to do it using that user data or another solution.