Scaling up and down on demand

Cognia hosts many of their applications in the Mendix cloud. Recently, they had a requirement to host one of their applications in their own public cloud account, while also leveraging the Mendix low code application development platform.

Mendix also supports public cloud providers such as AWS, Azure, and GCP, in addition to Mendix Cloud. Cognia approached EPI-USE to implement Mendix for private cloud (AWS) for their application. Since the application will be deployed in Cognia's own AWS account, they will have granular control over it. At the same time, the frontend and application development platforms can continue to use the current Mendix framework.

A successful migration and bespoke, scalable solution

To address Cognia’s requirement, EPI-USE Services for AWS designed a highly-available AWS infrastructure. It had the ability to scale up resources on demand to accommodate many concurrent users, and achieve optimum performance in peak periods. The infrastructure was also designed to scale down to minimum resources automatically during low-demand periods, thus optimizing costs while improving the application’s performance.

EPI-USE Services for AWS migrated the application to the AWS cloud and deployed it on the AWS-managed Kubernetes cluster, EKS. Amazon EKS is a fully-managed service that allows hosting Kubernetes on AWS while also handling the Kubernetes control plane. This simplifies the deployment, management, and scaling of containerized applications.

Instead of relying on the cluster autoscaler, using Karpenter helped improve the Cognia application availability and cluster efficiency by rapidly launching right-sized compute resources in response to changing application load. Karpenter offers more flexibility, and is well-suited for clusters with workloads that encounter periods of high, spiky demand, or have diverse compute requirements. When installed in the application cluster, Karpenter observes the aggregate resource requests of unscheduled pods and makes decisions to launch new nodes and terminate them to reduce scheduling latencies and infrastructure costs. Karpenter does this by observing events within the Kubernetes cluster and then sending commands to the underlying cloud provider’s compute service, such as Amazon EC2. 

Also, coupling Karpenter with horizontal pod autoscaling further improved the scaling process by deploying more pods with an increase in the workload. The Kubernetes Horizontal Pod Autoscaler automatically scales the number of pods in a deployment, replication controller, or replica set, based on that resource's CPU utilization. This helps the application to scale up to meet increased demand or scale down when resources are not needed.

To determine the appropriate infrastructure budget, EPI-USE Services for AWS conducted a load test using automated scripts to simulate actual traffic and assess the application's ability to scale up or down on demand. By analyzing the results, we determined the required resources to handle specific numbers of concurrent users, including the appropriate node type and costs for various traffic needs.

EPI-USE Services for AWS designed a solution for Cognia that achieved many benefits for the client:

• A highly-available AWS infrastructure
• The ability to scale resources on demand
• Achieving optimum performance in peak periods.