Background:
As we know that cost optimization is something, which helps in planning IT budgets for the organization and with the rising prevalence of AWS, I thought of writing about cost optimization to address the pricing challenges involved.
This blog will help you in reducing your AWS billing by 50 to 70%,
Always choose “Selective Utilization” modal instead of 100% utilization modal,
What is selective utilization modal ?
Let’s say – you are running m4.medium instance with 100% utilization modal,
Production: 24/hours day — $87.84/Month
QA: 24/hours day — $87.84/Month
Testing: 24/hours day — $87.84/Month
Development: 24/hours day — $87.84/Month
Total: $351.36/month
Selective Utilization
Production: 24/hours day — $87.84/month
QA: 12/hours day — $43.92/month
Testing: 12/hours day — $43.92/month
Development: 12/hours day — $43.92/month
Total: $219.6/month
If you shutting down the QA, Testing and Development bed’s during weekend’s it will further reduces your AWS bill, This savings could be substantially higher for larger or more resource-intensive environments, typically one can save 30 to 50 % on their billing here.
Instance Right sizing
Step 1 – Thumb rule here your resource utilization should be 80%, anything below this is called under utilized, so please make sure correct instance type is chosen and also keep a breaking point to monitor your instance utilization, you can also configure cloudwatch to find the statistic report.
Let’s say one of your production instance is running with the instance type : M4.Large instance with a ideal breakpoint of 6 months and its utilization is on a average of 80 %, then we have to move to step 2 – which is purchasing RI for it.
Step 2 – what is RI ?
Reserved Instances is nothing but providing commitment to AWS that we are going to utilize this instance for agreed time, which is 1 year to 3 year’s, choosing RI can provide you a significant discount (up to 75%) compared to On-Demand pricing.
Note: Always keep your instance virtualization type and instance family updated – new generation instance family will not only available with the lesser price it also having higher hardware configuration compare to what you have on older generation.
Eg – M1.Medium price is higher than that of M3.Medium
M3.Large price is higher than that of M4.Large
Account Separation & Tag creation
Always have seperate account for test and dev bed’s – that once have a keen picture how much we are spending for these also, we can have a account owner for accountability. Taging resources will also helps in identifying the places where we spend excess and one can work on that.
Cleanup orphan resources – Run periodic cloud housekeeping to check whether there is any left over orphan resources like snapshot’s, volumes, instance running unattended etc, one good example is cleaning up AMI won’t cleanup its associated snapshot and the leftover snapshot remain on S3 will continue to add up towards your monthly charges.
Follow our Twitter and Facebook feeds for new releases, updates, insightful posts and more.
