Manually verify flag-driven ASB cutover in test environments, incl. performance comparison

[martivj]

Description

Automated tests cover correctness in isolation; this issue covers behavior once everything is actually deployed and wired together in a test environment — flipping flags by hand, watching real queues drain, and comparing performance old vs. new.

Scope

• Manually flip each publish/listen flag pair (registration, inbound, outbound, validation) one at a time in the test environment; confirm messages flow through the expected path and in-flight messages on the old path still drain. Record findings/notes per flag.
• Run k6 performance test scripts against both the legacy Storage Queue + Functions path and the new ASB path for outbound and validation delivery (the two paths with a Functions hop either way), and compare latency/throughput for regressions.
• Document any regression found, with enough detail to file a follow-up bug if needed.

Acceptance Criteria

• Manual verification notes recorded for all four flag pairs.
• k6 performance comparison run and results documented (old vs. new path).
• Any regressions found are filed as follow-up issues.