Zscaler Digital Experience (ZDX): monitoring and troubleshooting
Measuring experience, not just blocking traffic
ZDX turns endpoint telemetry into per-user, per-application quality scores so you can answer the question 'is the problem the device, the network path, or the application?' The Client Connector collects metrics along the path: Wi-Fi signal and interference, DNS resolution time, TCP connect time, TLS handshake time, HTTP request and response time, page load, and - for private apps - ZPA performance. Cloud Path monitoring probes the legs from the device to the Zscaler service edge and from the service edge to the destination.
Scores and signals
ZDX produces a composite score (0–100) for devices and applications, derived from those metrics, plus availability indicators and configurable alerts. Use the score to triage, then drill into the underlying metric; a low score on its own is not a diagnosis.
Deep Trace
Deep Trace is the troubleshooting view: hop-by-hop path data with latency and packet loss at each hop, so you can localize degradation to the home network, the ISP, the path into the Zscaler cloud, or the cloud-to-application leg.
Traps
- Treating ZDX as a packet capture or a log/DLP tool - it is metric-based monitoring.
- Blaming Zscaler when Wi-Fi or the local ISP is the failing hop; the hop data shows it.
- Expecting ZDX data without the Client Connector deployed and enrolled.
- Comparing a device score with an application score as though they measured the same thing.
Rule of thumb
Many users at one location degrade at once → suspect the local path or ISP. One user degrades across all apps → suspect the device or Wi-Fi. One app degrades for everyone → suspect the application or its cloud leg.