Navigating AWS Complexity through EFS Cost Optimisation

AWS has grown increasingly complex, making it essential to grasp its finer details. Recently, I helped a client cut their Elastic File System (EFS) bill by $2,000 per month. They assumed this was due to their product’s growth as more users onboarded, but that wasn’t the full story.

The Cost of Blindly Following AWS Recommendations

AWS often provides configuration recommendations, but applying them without scrutiny can backfire. In this case, an AWS-recommended EFS setup was driving the 2000$/month high cost.

Understanding EFS Throughput Modes

Amazon EFS offers three throughput modes:

1. Bursting Throughput: The mode, where throughput scales with the amount of data stored. There are no extra charges beyond storage costs.

2. Provisioned Throughput: Allows you to set a fixed throughput level, independent of storage size. You pay for the provisioned capacity.

3. Elastic Throughput(recommeded by AWS): Automatically adjusts throughput based on workload. Apart from storage cost, you pay based on the amount of data read and written.

The client was using Elastic Throughput, recommended by AWS. However, their workload involved a high volume of data reads and writes, leading to steep costs.

Finding the Solution with CloudWatch Metrics

By analysing CloudWatch metrics for EFS, I noticed the client’s peak throughput was consistent and predictable with occasional erratic needs. Elastic Throughput was overkill for their needs and switching to Provisioned Throughput which can cover their occasional erratic needs made more sense. After making the switch, the client saved approximately $2000 monthly without any performance issue.

Key Lesson

AWS recommendations aren’t one-size-fits-all. Regularly monitor your usage patterns—CloudWatch is your friend here—and tailor configurations to your specific workload. A simple analysis can uncover significant savings.