I asked the following question on the AWS forum. I’ll update this post when I
get an answer.
AWS’s Spot Instance Advisor estimates the “frequency of interruption” for
each instance type in each region.
For example, In us-west-2, the t2.large currently has a frequency of
interruption of 5%, and the t3.large of 5-10%.
But what does “5%” actually mean?
I found one other reference to interruption percentages in AWS podcast “#305:
Saving Money with EC2 Spot”, published 2019-04-01.
In the last three months, 93% of the reason why spot instances got terminated
is because the user was finished with them and just turned them off. However,
they can also be terminated if the prevailing market price for the spot
instance exceeds your bid price.
So is Spot Advisor telling me that 5% of t2.large instances are interrupted
because of spot market movements?
Or is it that every hour that the t2.large instance runs, it has a 5% chance of
Any clarification on how to interpret these numbers would be greatly