For App Connectors, why shouldn't the customer pre-configure memory and CPU resources to accommodate a higher bandwidth capacity, like 1 Gbps or more?

While App Connectors require adequate CPU and memory, simply increasing these resources on a single instance (scaling up) does not linearly increase throughput to 1 Gbps or more. The primary bottleneck at such high traffic volumes becomes operating system-level limitations. A single App Connector must manage thousands of concurrent connections, and this can lead to the exhaustion of available ephemeral ports or file descriptors long before the CPU or memory are fully utilized. The recommended Zscaler architecture is to scale out by deploying multiple App Connectors in an App Connector Group.

A. While cost is a valid business consideration, it is not the primary technical reason for the architectural recommendation to scale out instead of up.

B. Storage I/O is generally not the primary performance bottleneck for an App Connector, which primarily functions as a real-time traffic forwarding proxy.

C. This is incorrect. Zscaler best practices explicitly advise against attempting to scale up a single App Connector indefinitely due to diminishing returns and OS-level bottlenecks.

1. Zscaler. (2024). App Connector Deployment Guide for VMware vSphere. Zscaler Help Portal. "To achieve higher throughput, deploy more App Connectors and configure them as part of the same App Connector Group. Scaling up an App Connector (i.e., adding more memory or CPUs) has diminishing returns."

2. Zscaler. (2024). About App Connectors. Zscaler Help Portal. "For high availability and to increase throughput, you should deploy App Connectors in pairs and configure them in an App Connector group." This highlights the scale-out model.