Introduction
HPE Aruba Networking User Experience Insight (UXI) is a powerful solution designed to provide IT teams with visibility into network and application performance from the end-user's perspective. The solution consists of two types of sensors:
Purpose-Built Sensors: These are physical devices positioned as close to end users as possible to emulate a user's network experience.
Agents: Software that runs on end-user devices within the network to monitor performance directly from their perspective.
Both types of sensors report results to the UXI cloud-based dashboard. The dashboard is the core feature of UXI, proving a simple to use interface allowing IT teams to:
View and analyze data collected from all sensors.
Filter and drill down into test results and issues for detailed troubleshooting.
Detect issues proactively from the end-user perspective.
This document explores how UXI helps IT teams detect and diagnose network and application issues, ensuring consistent and reliable user experiences.
Overview of the Sensor Test Cycle
UXI hardware sensors continuously monitor network performance by conducting synthetic tests across wired and wireless networks. These tests assess whether the sensor can connect to the network, obtain an IP address, DNS availability, external connectivity, captive portals and application performance. Sensors perform tests one at a time in a round-robin fashion.
UXI Agents gather network statistics, device details and run synthetic application tests. The key difference is agents run directly from the end user device in the background and do not disrupt what the user is doing. Therefore, agents do not test things like association or DHCP which would disrupt the user.
For more details, refer to the User Experience Insight Sensor Test Cycle - https://help.capenetworks.com/en/articles/3529400-user-experience-insight-sensor-test-cycle
Issue Detection
There are two types of issues observed in the dashboard that generate notifications & alerts:
Threshold Violation - If the sensor runs a test and gets a successful result, the test's measured value is reported to the dashboard. Threshold issues are generated based on a rolling average of measured data samples crossing a defined threshold. These thresholds are configurable by an administrator of the dashboard. Example issues: High DHCP response time, High DNS lookup time, High Application Latency.
Test Failure - If the sensor runs a test and does not get a successful result, the sensor will attempt to diagnose why the test failed in an automated triage mode. Triage mode will run a set of predefined troubleshooting tests to get the test failure's root cause. Example issues: Gateway is unreachable, Primary DNS Failing, No Response from DHCP Server, External Service Unavailable.
For a detailed explanation of issue types, visit Types of Dashboard Issues - https://help.capenetworks.com/en/articles/5137797-types-of-dashboard-issues
Diagnosing the Root Cause
UXI provides tools and workflows to help administrators efficiently isolate and resolve issues. Below are examples of how UXI facilitates root cause analysis:
Threshold Violation Issues:
Historical Trends: Analyze up to 30 days of historical data to identify recurring patterns of performance. Use filtering options to drill down by groups and networks for a more focused analysis.
Identify Spikes: Spot performance spikes at specific times of the day and correlate the spikes with network performance metrics such as RSSI, retry rates, channel utilization, which AP the sensor is connected to and more. Easily navigate between service pages to identify which sensors are reporting performance issues. From there, access the sensor status page to view the context of the measurements and gain deeper insights.
Path Analysis: Compare network paths during normal and problematic periods to identify anomalies. Leverage metadata such as location, DSCP markings, and AS numbers to pinpoint the root cause of the issue. Learn more about Path Analysis. https://help.capenetworks.com/en/articles/9426925-using-path-analysis-in-your-workflows
On-Demand Packet Capture: If the issue is ongoing, initiate an on-demand packet capture to validate the measurements and gain further insight into the problem.
Test Failures
When the sensors run a test and the test fails, the sensors enter an automated triage mode. The key features of triage are:
Triage does a step-by-step analysis to identify the root cause from an end user perspective. This helps narrow your focus to the root cause of the issue.
The triage menu contains the issue time, and if the issue is ongoing triage indicates when the issue was last confirmed by the sensor.
Expand the context of the test. This is helpful for identifying things like what AP the sensor was connected to, what channel the sensor was on and much more.
Expand and view the Linux-style outputs. This is especially useful for viewing failures such as DHCP, where you can see where the DORA did not complete, as well as application issues, where you can see the packet drops.
If the sensor is connected to both wi-fi and Ethernet, and the issue happens on wi-fi, triage will also run the tests on Ethernet to provide additional context so that one can one can easily determine if a certain issue is only failing on the wireless network or wired network, or both
In many cases in triage, one of the steps is to run a traceroute. This is useful for identifying the path from the sensor to the target at the time of the test.
Triage issues can be exported to plain text to share with colleagues.
Triage issues usually have a packet capture that can be used to validate the issue observed.
The help article further outlines how to troubleshoot with UXI sensors - https://help.capenetworks.com/en/articles/7967857-troubleshooting-guide
AIOPS Incident Detection
When you are managing hundreds or thousands of sensors, UXI AIOPS Incident Detection helps focus your attention on the most critical problems. With Incident detection enabled, UXI examines the issue data in real time to identify issues that are significantly different from typical issues seen in the customer environment. Issues are represented in blue. These anomalous issues identified by Incident Detection are consolidated into incidents and surfaced on the dashboard in red. An incident is a collection of related anomalous issues that share the same start time, networks, issue codes and services. More information here - https://help.capenetworks.com/en/articles/5132163-aiops-incident-detection
Conclusion
UXI streamlines the process of detecting and diagnosing network issues. By leveraging its advanced tools and actionable insights, businesses can ensure reliable network performance and superior user experience. This document highlights the platform's ability to not only pinpoint problems but also offer root cause analysis of the issue, making it a powerful resource for modern IT environments.