One of the 4 well architected pillars in AWS is Resilience. This translates to how your infrastructure performs without failures and interruptions.
In this article, we quickly summarize what makes a solution quickly reliable.
Below is the architecture of a reliable 3-tier infrastructure using EC2 instances distributed across three Availability Zones.
To practically do the lab, find it here
A simple implementation of the project would be comprised of an instance and the dynamo db table.
The service makes an API call from the dynamo db and serves it to the instance to be visible on a web server.
Best practices for reliability present
- The infrastructure contains an autoscaling group, with parameters having a minimum and maximum number of instances to scale to. This ensures that demand is met while keeping costs low with a maximum cap.
- Autoscaling groups also contain a feature where health checks of instances are performed and upon detection of a failure, unhealthy instance is replaced, automating the remediation process.
- Having the instances spread out over three availability zones increase reliability since the service is less prone to down-time. When one zone isnʼt available, traffic is directed to the instances that are up.
- The elastic load balancers distribute traffic to the instances to help alleviate stress on an instance while others are less busy. This also decouples the infrastructure as the service is not dependent or fixated on a particular instance.
Top comments (0)