Section8: High Availability and scaling

Question 1

Why use Elastic Load Balancer?

Answer 1

• Its a managed load balancer, AWS takes care of it.

Question 2

Name 4 types of load balancer on AWS

Answer 2

1) Classic Load Balancer(v1 - old generation) 2009
- Supports Http(layer7),Https TCP(layer4),SSL
2) Application Load Balancer (v2 - new generation,Layer7 load balancer) 2016
- Can do routing
- Are great for micro services and container-based applications
- Supports HTTP, HTTPS, WebSocket
3) Network Load Balancer(v2 - new generation,layer 4) 2017
- Support TCP,TLS,UDP
4) Gateway Load Balancer 2020
- Operates at layer 3(Network layer)

Question 3

What is SNI?

Answer 3

• Stands for Server Name Indication.
• Solves the problem of loading multiple SSL certificates onto one web server.
• Its a “newer” protocol and requires the client to indicate the hostname of the target server in the initial SSL handshake.
• The server then find the correct certificate or return the default one.
• Note:
  - Only works for ALB & NLB,CloudFront
  - Does not work for CLB.

Question 4

What is Connection Draining?

Answer 4

• In CLB it is called Connection Draining but in ALB & NLB it is called Deregistration Delay.
• Time to complete “in-flight requests” while the instance is de-registering or unhealthy.
• Stops sending new request to the EC2 instance which is de-registering.
• Time is set between 1 to 3600 seconds(default:300 seconds)

Question 5

Things to remember about ASG(Auto scaling groups)?

Answer 5

• Scaling policies can be on CPU, Netowrk or custom metrics.
• ASG use Launch configurations or Launch Templates
• To update an ASG, you must provide a new launch configuration/launch template.
• IAM roles attached to an ASG will get assigned to EC2 instances.
• ASG are free, you only pay for the resources used.
• Having instances under the ASG means that if they get terminated for whatever reason, the ASG will automatically create new ones as a replacement. Extras safety.
• ASG can terminate instances marked as unhealthy by an LG

Question 6

Auto Scaling Groups - Dynamic scaling Policies

Answer 6

1) Target Tracking Scaling
- Most simple and easy to setup
- Example: I want the average ASG CPU to stay at 40%
2) Simple/Step Scaling
- When a CloudWatch alarm is triggered(example CPU > 70%) then add 2 units
- When a CloudWatch alarm triggered(example CPU < 30%) then remove 1 unit
3) Scheduled Actions
- Example: increase the min capacity to 10 at 5pm on Fridays
4) Predicting scaling
- Continuously forecast making use of machine learning

Question 7

Good metrics to scale on for Auto Scaling Group?

Answer 7

• CPUUtilization
• RequestCounterPerTarget
• Average Network In/ Out
• Any custom metric