Description of problem:

I am scale testing BGP router export scenario on a OCP 4.19 baremetal cluster with 24 worker nodes (no infra nodes) using 4.19.0-ec.3 with  OVNK BGP image which is built using the PR  build 4.19,openshift/ovn-kubernetes#2239 on 03/25/2025.

Test creates 72 CUDN, 1 namespace per CUDN and 1 pod per namespace. I waited for 30 minutes after creating all the resources and before creating router advertisements.

Then test creates 72 router advertisements where each RA advertises only one unique cudn i.e RA:CUDN is 1:1.

We are monitoring on the baremetal host where the external frr is running for the route updates using netlink sockets. Latency of Router advertisement is the time elapsed from the creation of the RA CRD to the detection of the corresponding route in this external frr host.

Latency of exporting RA to external frr linearly increasing with the number of RAs and taking more than 5 minutes for 70th RA. 

Another observation is it is hardly taking 2 seconds for the ping test from this external host (once it detects the route) to the pod inside the cudn.

Test results are avaialble at https://docs.google.com/spreadsheets/d/1WLuTpcrTwFBUcZ-XF2wppOJL9M13_V_HYh2i23UpQUo/edit?usp=sharing 
Grafana screenshots at https://storage.scalelab.redhat.com/anilvenkata/bgp/ra72export/ 

I can provide the live environment to the engineer for troubleshooting.