DynamoDB: How is read/write capacity calculated and what does it mean (for the free tier) if it is scaled quickly?

Question

So I have been working on a couple Lambda functions for a Alexa skills including DynamoDB (actually 5 tables each with 5 RCU/WCU. I am on the free tier and I haven't used the work-in-progress functions for a couple of days, but today I realized in the cost overview that I have used about 9000 hours and will possibly use about 12000 at the end of the month. See here:

So was wondering why? I know that you reserve the hourly capacity. And since none of the skills is live, I sticked with 5 RCU/WCU for each table (it says 25 RCU/WCU is okay for the free tier). But these numbers doesn't make sense?

Or do they? And the question is what happens when I scale this down to 1 RCU/WCU when all the skills go live?

I know that this would mean that only one request per second can be processed, but I am sure that this would be enough for the skills (at least for the first couple of month) - I guess that I won't have more than 200-300 monthly users. Will this significantly reduce the above numbers? Or will I be charged huge bill for some reason I don't understand after going though the documentation?

Answer 1

The numbers do make sense, and yet you appear to be safely within the free tier.

The problem is in understanding the dimensions.

The page you are viewing shows capacity unit hours.

25 CU-hours per hour × 744 hours per month ≈ 18,600 CU-hours per month of each type (read, write) that are allowed in the free tier.

This number (18,600) is consistent with the percentage calculations shown: 12151 ÷ 18600 ≈ 65%, as indicated.

and will possibly use about 12000 at the end of the month

Slightly more than that. You need to add the regional numbers together, so 12151 + 2177 is your projected total, which is ~78% of the total allowed.

You don't appear to need to scale anything down at this point.