Avoiding be billed: can I replace t2.xlarge by t2.micro?

I created a t2.xlarge instance that uses 30 Gb. I noticed that with t2.xlarge I am being billed because the instance is out of Free Tier configuration. For this course, can I change from t2.xlarge to t2.micro and continue to use the same 30Gb volume? T2.micro is enough for this course? Yes, I still need create a p2.xlarge instance to be able to use gpu to compute all the datasets of this course. In short words, I am thinking to have a t2.micro, p2.xlarge and a 30Gb volume shared for both instances. Is that feasible?


Jeremy suggests the approach of getting your tests to run on t2.micro on sample data and then if you feel it’s necessary to get a quicker result spinning up a larger instance / gpu instance to run your full experiment there.

1 Like

Hi @glyph, this approach sounds good. In this case I can use t2.micro or even my local computer to run the samples and then, if I want to get better idea about the full dataset or even if I want to send my results to Kaggle, then a gpu instance is the way and I will be billed for just one instance instead of two.

1 Like