For Updates testing we are doing a validation with continuous pings of the workloads and we track how many were lost, allowing us to fail the job if the lost ping count goes over some threshold. We should implement such functionality in the Adoption test suite as well.

Acceptance criteria (draft):

• Add a continuous ping test to at least one of the Adoption jobs that are deployed with three controllers:
  • Preferred one is periodic-internal-adoption-multinode-to-crc-ceph, alternative is periodic-internal-adoption-multinode-to-crc-non-ceph.
• The ping test starts before the Adoption procedure and finishes after the Adoption procedure.
• Excessive ping loss makes the job fail. We can set in a variable how much ping loss is considered excessive.
• The ping test should be emulating at least N-S traffic on a floating IP.

---- * Stretch goals that can be split off into separate epics if they aren't easy to achieve:

• • E-W traffic ping test.
  • Adding the ping test into one of the upstream jobs (periodic-adoption-multinode-*).

Note: The Networking DFG already has their ping test hooks in the test suite: https://github.com/openstack-k8s-operators/data-plane-adoption/pull/582/files . This or something similar would be a valid approach, but so would be using test-operator, as was mentioned in the preliminary discussions.