Introduction

We were trying to "measure" the cost of usage of a specific use case on one of our Aurora DBs that was not used otherwise.

2020 Oct 8 between 19:18 and 19:19 hrs. UTC we issued some representative queries to it. This was 1 minute (!) of action.

Then, the next day, we went to the CloudWatch Insights and selected the graphs for read/write IOs for the day before like in the graph below:

Observation

From the graph it is suggested that there was a constant read and write IO load between 18:50 and 19:50. However, our little experiment took place at 19:18 (yes, exactly at half time). Our experiment took only 1 minute, but the graph suggests a whole hour of activity.

Questions

First, why is there not a single point at 19:20 hrs.? Instead, the line is up for a whole hour, exactly around the time our our action. Artefact?

Second, and more importantly, if that is an artefact, will we be billed for 12 times the read/write IOs or for only the 1 at 255?

Note: This question was spun off of this one: AWS CloudWatch interpreting insights graph -- how many read/write IOs will be billed?