An upgrade to a web application, which supports 400 users at four sites, is being tested. The application runs on four servers behind a load balancer. The following test plan is proposed: - Have 50 us

Sizing tests whether servers have sufficient CPU, memory, and storage capacity to handle expected workloads - not tested here since only 50 users per server are used and capacity limits are not being probed.

By assigning users from each geographic site to a specific backend server and bypassing the load balancer, the plan validates that end-to-end network connectivity exists between each site and each server. Connectivity testing confirms that routing, firewall rules, and network paths allow traffic to flow from source to destination, regardless of load or failover behavior.

High availability testing requires simulating component failures - such as taking a server offline - to verify that the load balancer redirects traffic and users remain operational.

Performance testing measures response times, throughput, and latency under realistic or peak load conditions, which this plan does not address since it focuses on path reachability rather than load metrics.