ELB

Rationale

AWS ELB  (Elastic Load Balancing) is the service we use for exposing applications to the Internet . It provides load balancers using an IaaS  model.

The main reasons why we chose it over other alternatives are the following:

• It seamlessly integrates with VPC , EC2 , EKS , etc., allowing for the easy serving of applications hosted in the cloud .
• It complies with several  certifications from ISO  and CSA . Many of these certifications are focused on ensuring that the entity follows best practices regarding secure cloud-based environments and information security.
• When combined with Kubernetes , it allows for balancing application load by distributing requests to multiple replicas  using a horizontal scaling approach .
• It has its own Kubernetes module  for automatically provisioning application load balancers  when Kubernetes applications are deployed. This is especially useful for serving ephemeral environments .
• It supports VPC security groups , allowing us to easily set networking inbound and outbound rules for the load balancers. Such a feature is essential for avoiding CDN bypassing .
• A single load balancer supports multiple availability zones , granting networking redundancy, which is essential for keeping it always available to the Internet .
• It supports health checks , allowing for the constant monitoring of all the endpoints associated with a load balancer. Application requests are only sent to healthy endpoints.
• Application load balancers  support rules , allowing us to create complex routing scenarios when it comes to request forwarding.
• It supports application load balancers, network load balancers , and gateway load balancers , providing infrastructure for a wide range of solutions.
• Load balancers can be monitored via CloudWatch .

Alternatives

Usage

We use ELB  for

• serving our platform production environment , and
• serving our ARM ephemeral environments .

We do not use ELB for serving our website , as it is a static site served by S3 , which directly provides endpoints  without having to manage load balancers.