KB40204 - Pulse Connect Secure Health Check and System Monitoring Guide

This article provides an overview of the system dashboard graphs and provides thresholds that PCS admins should be aware of when monitoring the health of the system.  


 

PCS admins can monitor the overall health of the system by navigating to System > Overview.  The dashboard graphs displayed on this page can provide important indicators of the overall system health and should be provided for any support case that is open for system performance issues, in the form of screenshots.



 

Pulse Secure recommends that the following thresholds should not be exceeded in the following dashboard graphs:

• CPU and Virtual Swap Memory Utilization graph:
  • CPU activity will go up and down depending on the features used and the number of concurrent users signed into the device.  This is normal.
  • CPU activity where the graph line remains above 80% in a straight line for several hours, or days, is not normal and may indicate that a process is leaking memory.  (Refer to KB15832 - Troubleshooting and recommended logs to identify high or 100% CPU issues). 
• Virtual (Swap) Memory Utilization:
  • Virtual Swap Memory graph only tracks virtual memory in use after all physical memory has been consumed.  If any swap memory is being used, this should be investigated further by Pulse Secure support team as it could indicate a memory leak.
  • Customers should monitor for periods of swap memory being in use in the graph and take a system snapshot while the swap memory line is above '0' and provide this when opening a case with Pulse Secure support.

• Concurrent User graph - The concurrent user graph line on this graph includes VPN tunnel users and all other users, such as Web, Terminal Services, File browsing users, etc.  
  • Can be monitored for any sudden drops in the number of users signed in which can indicate an outage has occurred that has caused sessions to be terminated suddenly.
  • Can also be monitored to insure that the number of concurrent users is not exceeding the license count allowed for the device. 
• Throughput:
  • Graph activity is measured in Megabits with an 'M' displayed on the 'y axis' or in Megabytes with 'MB' displayed on the 'y axis'. 
  • The threshold for throughput will depend on a number of different factors:
    • ESP versus SSL mode
    • Packet Length (Larger payload volumes will increase throughput and smaller payloads will decrease throughput)
    • Hardware form (Each PSA and MAG models will have different thresholds)
  • In general, if the throughput is exceeded, performance will be impacted and considerations should be made to split the load between two Active/Active nodes or upgrade the hardware to support the load. 
  • External and Internal port throughput should be similar and the graph lines for each should line up fairly closely.  If the external throughput differs greatly from the internal throughput, or vice versa, then this could indicate a large number of packets are being dropped.  

​Other statistics to monitor on the Status page that are not graphed:

• State Storage:
  • Should always be less than 60% full and not exceed 50% fragmented at any given time.

​

• ​Logging Disk: 

• Percentage should be below 80%.  Performance issues can occur as the logging disk reaches 100% full and should be maintained by deleting system logs/snapshots that can be recycled.

Customer can connfigure SNMP to monitor the health of the device, this has more description of the device resources. Refer to PSC-SNMP-Monitoring-Guide.pdf
Login to the Admin page to view the graph below. 
 

SNMP can be used to monitor the health of the PCS device.  SNMP provides more details of resource usage and can be configured to alert a PCS admin if thresholds are met.

Refer to Configure SNMP Monitoring Guide.

Related Knowledge Base articles for PCS appliance monitoring.  
 
KB26207 - SNMP monitoring of PCS, MAG PCS, MAG PPS and PPS devices
KB12276 - What is the download location of the Private MIB file for PCS/PPS device?
KB20715 - Cannot obtain values of some SNMP MIB entries
KB25292 - How to monitor hard disk usage
KB14263 - Best practice recommendation: Monitoring memory utilization via SNMP
KB22201 - How to view state storage statistics
KB20495 - How to recover from "Persistent storage is full - Unable to recover system state"
KB15832 - Troubleshooting and recommended logs to identify high or 100% CPU issues
KB17044 - "Facility" in Syslog Servers window under Log/Monitoring tabs: What does it mean?
KB24954 - Pulse Connect Secure / Pulse Policy Secure is sending 'iveLogNearlyFull' SNMP traps