Users are reporting CPU related performance problems on the virtual machine Email-Prod throughout the day. The resource settings for the VM and the ESXi host the VM is running on are shown below: VM:

Numa.PageMigEnable only controls whether page migration between NUMA nodes is permitted at all - enabling it without configuring memory affinity does not directly optimize placement for this VM's workload.

Numa.AutoMemAffinity instructs the hypervisor to automatically migrate memory pages to the NUMA node local to the executing vCPUs, reducing costly remote memory accesses that occur when the VM spans both NUMA nodes.

Setting CPU affinity to pin 2 vCPUs per NUMA node aligns each set of vCPUs with a local memory domain, allowing each group to access memory from its own NUMA node and avoiding the latency penalty of cross-node memory traffic.

Pinning all 4 vCPUs to a single NUMA node concentrates all memory and CPU pressure on one node, which can increase local contention and does not leverage the capacity of the second node.