Unable to scale Azure cluster to 120 nodes without failed machines and nodes never become ready

NOTE: this functionality passed on 4.16 
And passed on 4.17 with a workaround documented here: https://access.redhat.com/solutions/6982343

Is this workaround a new every day change for scaling above 55 nodes?

Machine events show: 

InvalidConfiguration: failed to reconcile machine "scaleci18-8929-xd67z-worker-centralus2-2k44r": network.InterfacesClient#CreateOrUpdate: Failure sending request: StatusCode=400 -- Original Error: Code="SpecifiedAllocatedOutboundPortsForOutboundRuleExceedsTotalNumberOfAvailablePorts" Message="Specified Allocated Outbound Ports 1024 for Outbound Rule /subscriptions/***/resourceGroups/scaleci18-8929-xd67z-rg/providers/Microsoft.Network/loadBalancers/scaleci18-8929-xd67z/outboundRules/OutboundNATAllProtocols exceeds total number of available ports per backend instance of 1008 based upon desired pool size. Reduce allocated ports or increase number of IP addresses for outbound rule." Details=[]
    

 4.18.0-0.nightly-2024-10-08-033830
    

100%
    

    1. Create azure ovn cluster on  4.18.0-0.nightly-2024-10-08-033830
    2. Scale all 3 zone machinesets (40 nodes on each of the 3 machinesets)

Can be done in the console or below command executed on each of the 3 machinesets

 oc scale machinesets -n openshift-machine-api ${machineset} --replicas 40
    3. Many failed machines and nodes don't become available 
    

Many machines fail and nodes don't become available 
    

All machines and nodes become running 
    

Did not hit this issue on 4.17.0-nightly**, this is the first time hitting this issue on 4.18. Running on different regions and newer versions of 4.17 for comparison of results 


Must gather can be found here: https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/must-gather/594/artifact/