Auto Scaling Solutions

Auto Scaling Solutions

• Auto scaling overview
  • It monitors applications
  • Adjust capacity ( add/remove depending on the need)
  • Manage costs
  • Scalable AWS resources
    • EC2 Auto Scaling groups
    • Aurora DB clusters
    • DynamoDB global secondary indexes
    • DynamoDB tables
    • Elastic Container Services (ECS) services
    • Spot Fleet requests
  • Auto Scaling Costs
    • Free to use
    • Results of use may cost
      • More instances
      • CloudWatch
      • ELB load balancer
• Auto Scaling groups
  • Collection of instances with similar characteristics
    • Can be scalable based on criteria
    • Unhealthy instances can be auto-replaced
      • Any state other than “Running” is unhealthy
  • Group Considerations
    • Time to launch and configure a server
    • Relevant metrics to your application. Scaling out based on:
      • CPU utilization
      • Network throughput
      • Free memory
    • What AZs should the Auto Scaling group span?
    • Scale to increase or decrease capacity?
    • Specify min number of instances always running
  • Termination policies
    • Scaling methods
      • Scaling out – adding instances
      • Scaling in – removing instances
    • You can define custom termination policies based on:
      • Oldest instance
      • Newest instance
      • Oldest launch configuration – take which has the oldest configuration
      • Closet to next instance hour – instance that about to be billed again (good for saving cost)
      • Default – according to the default policy
• Load balancing concept click here for more details
  • Definition: Need a way to have multiple servers to perform the same operations
• Elastic Load Balancing (ELB)
  • ELB uses Dynamic load balancing not static
  • ELB benefits
    • Highly available
    • Secure
    • Flexible
    • Monitoring and auditing included
    • Elastic – can have multiple ELB
    • Hybrid – implement multiple types of load balancing within ELB
  • Types of ELB
    • Application load balancer
      • Good for HTTP and HTTPS traffic
      • Provides advanced request routing targeted at the delivery of modern application architectures which includes
        • Microservices
        • Containers
      • Operating at the individual request level (OSI layer 7)
      • Routes traffic to targets within Amazon VPC
    • Network load balancer
      • Best for load balancing TCP traffic
      • Operating at the connection/transport level (layer 4)
      • Capable of sudden and volatile traffic patterns
    • Classic load balancer
      • Oldest model (rarely use)
      • Only for legacy use, not for new deployment
      • Basic load balancing across multiple Amazon EC2 instances
      • Operates at both request and connection level

Note: here is the link for more ELB comparison https://aws.amazon.com/elasticloadbalancing/features/

• Supported services are:
  • EC2 – virtual machines
  • ECS – dockers
  • Auto Scaling – monitors and adjust application accordingly
  • CloudWatch – monitoring and management service
  • Route 53 – DNS
• Virtual Network Services
  • Domain Name System (DNS) overview click here
  • AWS Route 53
    • It’s a DNS service provided by AWS
    • It’s a managed service (which means you don’t have to create an EC2 instance)
    • Route 53 routing policies
      • Simple – simply resolve domain name with the IP address
      • Weighted – route to multiple servers
        • For example, you want to route 80% of traffic to IP 192.168.10.1 and 20% to 192.168.20.1
      • Latency – manually sending to the user to closest destination for lowest latency
      • Failover – only use another instance if one instance. It is the same as using an active/passive model of two servers
      • Geolocation – it routes traffic based on users originated location so user can have less latency
        • For example, Route 53 gets a user’s geolocation and forward them to a closet server
      • Multivalue Answer – evaluate the health check of multiple instances
        • Multiple health checks up 8
        • It is not a substitute for a load balancer

Note: more details about Choosing a routing policy https://docs.aws.amazon.com/Route53/latest/DeveloperGuide/routing-policy.html

• Flow logs
  • It allows you to capture traffic in your AWS account
  • How it works
    • Copy your Bucket ARN number from S3 buckets where you want to save logs
    • It monitors traffic on:
      • EC2 instance
      • Subnets
      • VPC
    • It stores that information in S3 buckets