2

We are researching on the possibilities of moving our app from EC2 based implementation to AWS lambda. We are trying to project how much money we will be saving if we do so. The lambda pricing says it gives 400000 GB-seconds free per month and a tiny amount for each GB-second after. How do I know from my AWS console, how much GB-seconds I am utilizing per month now?

user3360140
  • 281
  • 1
  • 4
  • 13
  • 1
    You can't tell from the console. You will probably have to look at how many requests go to your application and how much CPU / RAM they need and estimate from there. – Tim Sep 29 '19 at 08:16
  • It will help if you post a screenshot of your utilization plots from cloudwatch (CPU/RAM if available). – Shadi Oct 02 '19 at 14:42

1 Answers1

0

If your utilization of the EC2 server is a spiky one (i.e. long times of inactivity and then a sudden spike followed by inactivity again), then it would be a good idea to move to lambda functions. A faster cost saver would be to move to a smaller burstable machine type (eg t2). These types let your CPU go beyond 100% for a period of time. Also, staying within the EC2 space means that you don't need to worry about changing your application architecture/deployment method to move to lambda.

A quick way to check if it's worth it to move to a burstable EC2 is the tool isitfit. You could install it with pip3 install isitfit and then just run isitfit --optimize to identify EC2 servers that are convertible to burstable type. These would also be convertible to lambda. You could filter the output for just the EC2 server that you mentioned.

Disclaimer: I'm the author of isitfit.

Shadi
  • 121
  • 5