How we helped achieve a $100K/year saving in cloud costs by switching cloud providers
Introduction
In the rapidly evolving landscape of cloud computing, optimizing cost while maintaining efficiency is a key challenge for businesses. This case study explores how one of our clients achieved significant cost savings by migrating dozens of their workloads from AWS Elastic Kubernetes Service (EKS) with Fargate clusters to Google Cloud Run instances.
Cost Savings: From AWS EKS Fargate to Google Cloud Run
Assessing Client Workloads
Our journey began with a thorough assessment of the client's workloads running on AWS EKS Fargate. The goal was to identify cost optimization opportunities without compromising on performance and scalability.
Auto-scaling to Zero with Cloud Run
A pivotal factor in achieving cost savings was Cloud Run's ability to auto-scale to zero. This feature ensures that when the services are not in use, they do not incur costs, unlike AWS EKS Fargate clusters, which do not scale down to zero. This capability of Cloud Run was crucial for our client, who had variable workload demands.
Managing Cold Starts in Cloud Run
While migrating to Cloud Run, we encountered the challenge of cold start times. Cloud Run instances, when scaled to zero, take some time to spin up when a new request comes in. However, the cold start time was within acceptable limits for the client's use case. We implemented best practices to minimize the impact, ensuring that the overall performance remained robust.
Deployment and Testing
Migrating Services to Cloud Run
Each service from the AWS environment was meticulously mapped and redeployed as Cloud Run instances. This process involved careful planning to ensure compatibility and performance optimization in the new environment.
DNS Setup and Integration
Post-migration, we set up the DNS to point to the newly deployed Cloud Run services. This step was critical to ensure seamless access to the services without any disruption to the end-users.
Extensive Testing
To guarantee the reliability and efficiency of the migrated services, we conducted extensive tests. These tests covered various aspects, including load testing, security vulnerabilities, and performance under different scenarios.
Conclusion
The migration from AWS EKS Fargate to Google Cloud Run presented a strategic opportunity for our client to optimize costs significantly. The ability of Cloud Run to scale to zero played a crucial role in achieving these savings. Despite the challenge of cold starts, the overall performance and efficiency of the client's services were maintained, if not enhanced. This case study exemplifies how careful analysis and strategic migration to the right cloud services can result in substantial cost benefits.
This blog post is intended for informational purposes and reflects the specific scenarios and requirements of the mentioned client. Results may vary based on different use cases and configurations.