I currently have an ec2 t2.micro instance and rds db.t3.micro instance on AWS, and in order to reduce costs I wish to transfer both the application (which is currently on the t2.micro ec2 instance) and database onto t3.nano or t3.micro instances. Thus my question is how do I analyse my current infrastructure in order to determine whether or not t3.nano/micro instances can handle the workload that my previous instances are handling? Do I look at the CPU percentage or something else? The CPU for my RDS is showing 4.75% at the moment - can I take this to mean my rds is using very little and thus could probably be shifted to something smaller?

CodePudding user response：

Note that T2 and T3 are Burstable performance instances meaning that they are very powerful if used in bursts, but provide lower performance when usage is sustained.

So, if your system becomes popular and has a continual load, the app/database might suddenly drop in performance when the burst credits run out. This won't be obvious with normal monitoring, which would show everything performing very when there are only occasional bursts of traffic.

Reducing the instance size will also reduce memory, which could cause problems with your app/database.

If you just have a simple app & database requirement, you could consider using Amazon Lightsail, which can offer considerable savings since it includes a free bandwidth allowance.