Hardware Health

Each of the following conditions triggers a hardware health anomaly.

Table 1. Anomalous hardware conditions
Condition Threshold
Non-user initiated reboots More than 5 in 1 hour
Non-user initiated reboots More than 10 in 24 hours
CPU utilization Average CPU usage greater than 90% for 10 minutes
Memory usage Average memory usage greater than 90% for 10 minutes

To open the Hardware Anomaly panel, select a device, for which anomalies have been detected, from the table. The panel includes the following information:

Header

This section provides general information about the anomaly. To get more information about the device, select Go to Device.

Click to expand in new window
Hardware Health anomaly header
Issue
This section identifies the condition for which the threshold was met, for example:

<name of device> memory utilization is consistently above 90%.

Recommended Actions
This section provides steps to resolve the anomaly, for example:

Identify the process or service causing high memory usage using the CLI command 'show memory detail'.

Time Series Analysis
This section presents relevant synchronized time series graphs. Expand the section to see the graphs. To choose which graphs to display, select Select Additional Charts.
Table 2. Additional charts
Chart Display on hover Expand on select
Total Throughput Total throughput Detailed chart showing: total, unicast tx, rx, broadcast tx, rx and multicast tx, rx traffic
Total Client Count Total number of clients Client details: MAC, client hostname, client OS
Reboot Reboot code and reason None
PoE Mode AP PoE power mode None
Device Temperature AP temperature None
CPU Utilization Percentage of CPU utilization CLI output for show system process state
Memory Utilization Percentage of memory utilization CLI output for show memory detail