in sales in 2021
Lack of Control
All the client's sites were hosted in the data center. Communication with the vendor was built through the ticket system. The waiting time for an answer or problem resolution could last days or weeks, no matter how big or small the problem was.
The client's infrastructure setup with separate servers for sites, its staging environment, database, and database staging led to scalability issues. Since only vertical scaling was possible, handling the increased workload and traffic demands took a lot of work.
Even though the client operates the business in one country and could turn the difference in day and night traffic to their advantage, it wasn't the case. With a regular server, the team paid for max capacity no matter the actual numbers.
Transferring from the data center to the cloud helped the client gain complete control over their infrastructure. Now they can manage, tweak, expand, and adapt any feature to their needs anyhow and anytime.
Minding client's need to unlock horizontal scalability, our team suggested switching to AWS Elastic Kubernetes Service. We complemented this solution by making the application stateless, storing client sessions in Elasticache for Redis, and streamlining CI/CD deployment from Bitbucket to the Kubernetes cluster.
Switching from on-demand to AWS spot instances helped to optimize the client's expenses (since its price can be up to 70% lower). To make that possible, our team made the system dynamic and configured it to finalize any requests in 2 minutes or less. So switching from one spot instance to another won't be a problem.