Just wrapped up a 30-day PoC for Zscaler Private Access (ZPA) and it was a no-go for us. The core issue? Unpredictable latency spikes that made our finance team's daily RDP sessions to their critical accounting server unusable. They'd get booted mid-transaction, which is a complete deal-breaker.
We followed Zscaler's recommended setup and had them involved. During demos and light testing, everything was smooth. But under real daily load—around 10 AM when the team is all online—the latency would jump from 30ms to 200ms+ for minutes at a stretch. The app connector was hosted in our own Azure region, so the issue wasn't our data center.
Here’s what we dug into during the evaluation:
* **TCO beyond list price:** The per-user cost seemed okay, but the required "Zscaler Client Connector" for every device added significant management overhead our team didn't anticipate.
* **Hidden hiccups:** The finance app isn't chatty, but it's sensitive to latency. ZPA's "closest-to-the-user" routing seems to introduce variability when the nearest PoP is under load.
* **SLA vs. Reality:** Their uptime SLA is solid, but it doesn't cover performance degradation like these latency spikes. For RDP and VDI, that's what actually matters.
Has anyone else faced similar performance issues with ZPA, specifically for latency-sensitive protocols like RDP or database clients? Did Zscaler provide any tuning parameters, or did you just have to architect around it? We're now looking at alternatives, and real-world performance data is gold.
buy smart
buy smart