How to Reduce AWS Costs Part-2
Gyan prakash
Vice President - Solutions Architect and AI agent Builder at J.P. Morgan, Snowflake, CrewAI, Autogen, LangChain, Author, Kubernetes, AWS, Azure
In the event that your association is engaged with an IT advancement on Amazon’s Elastic Compute Cloud (EC2) administration, discovering methods of how to decrease AWS expenses can be an issue. Normally you would prefer not to restrict the assets your engineers have accessible to them, yet you likewise need to ensure that the activities they are chipping away at stay inside spending plan and are monetarily practical.
Major sources of wasted cloud spend
Idle Resources: non-production resources, such as those used for development, testing, staging, and QA, are typically only needed during a 40 hour work week. Since AWS on-demand charges are based on the amount of time that resources are left running, that means that on nights and weekends (about 65% of the week), the spend on these resources is wasted.
Oversized Resources: AWS offers a number of sizes for every instance options, and for many users, the default behavior is to choose the largest size available. After all, you don’t know what capacity you might need in the future, right? However, a recent study performed by ParkMyCloud found that the average utilization of deployed AWS resources was a mere 2% — indicating routine overdeployment. By reducing an instance by one size, you reduce the cost by 50%. Reducing by two sizes saves 75%.
Cutting spend on resources that are completely unnecessary is the easiest way to quickly and significantly reduce AWS costs.
By How Much Can You Reduce AWS Costs?
Cost reductions through scheduling are easy to estimate. Based on a pre-configured schedule of running non-production EC2 and RDS instances only between 8.00 a.m. and 8.00 p.m. Monday to Friday, you will reduce AWS costs by 64%. Your AWS cost reduction on non-production instances could be higher or lower depending on the schedules you actually apply and the number of times they are interrupted outside those hours as the table below demonstrates.
Further Ways to Achieve an AWS Cost Reduction
Leaving non-production assets running is one of the biggest contributors to “cloud waste” — estimated to represent around 30% of all cloud spend — and whereas scheduling stop/start times for non-production instances will eliminate some of the waste, it will not eliminate all of it. This is because assets are often deployed in the cloud with more CPU, memory and storage than they need, or because demand decreases over time, or because assets are left active when they are no longer required.