metrics: Start to build a set of tests to track progress/regressions

[grahamwhaley]

This is a placeholder/high level view issue. Expect sub-issues to implement/fix/add specific items.

We should start to build up a set of metrics tests that we can run (both in a regular/automated manner via a CI and by hand if necessary) to monitor progress and/or regressions in the stack performance, both due to stack changes and/or underlying OS updates.

We can start with the basic measurements:

• pod launch times
• pod footprints
• pod scaling limits
• cpu overheads

and then move on to more complex testing:

• pod response times and availability
• inter-pod container testing
• cross-pod testing
• both storage (volume) and network tests

[grahamwhaley]

/cc @krsna1729 @mcastelino @obedmr @dklyle @kad

The idea is we will make a new subdir for the tree of tests and support tools.

[mcastelino]

@grahamwhaley please do. This makes perfect sense.

[knangia]

Closing this, since the PR is merged.

[jascott1]

Our bad, re-opening.