Description of problem:

We see ~40% increase in P99 pod ready latency and ~30% increase in ovn cpu utilization after the PR https://github.com/openshift/ovn-kubernetes/pull/2417 was merged.
The increase can be seen starting from 4.19.0-0.nightly-2025-01-22-232832

How reproducible:
Always

Steps to Reproduce:

1. Run udn-density workload and observe

Actual results:

Worse than previous runs

Expected results:

In line or better than previous runs