Summarizing a corridor talk I've had with @fruch and @yaronkaikov :
The idea is to ensure a new AMI (With a new Linux kernel, or some backports) doesn't introduce a very obvious performance regression.
We should run a minimal (small) performance tests on a fixed (small) setup and see that we did not regress considerably.
Example:
Summarizing a corridor talk I've had with @fruch and @yaronkaikov : The idea is to ensure a new AMI (With a new Linux kernel, or some backports) doesn't introduce a very obvious performance regression. We should run a minimal (small) performance tests on a fixed (small) setup and see that we did not regress considerably. Example: