Test for Zero Downtime Scaling of NGF pods

[pleshakov]

Consider two cases of horizontal scaling of NGF pods:

• scaling up
• scaling down

For both cases, it is critical that external clients don't experience any errors (downtime) when sending/receiving traffic to/from the applications exposed via NGF.

Acceptance criteria:

• Determine whether scaling up or down results into any errors seen by clients.

[ja20222]

Blocked by #950

[pleshakov]

Hi @brianehlert @jasonwilliams14

What is the maximum number of NIC replicas you saw among users/customers?

This will help us determine the max number of replicas to scale up in this test.

[brianehlert]

I do not recall many beyond 25. I am sure there are, as there are customers who dedicate clusters to ingress exclusively.
But probably a reasonable place for now.