Fault Monitoring on Windows Integrity Servers

Methodology
Foundation Agents collect data at the data collection interval set in the control panel applet,
(default is 2 minutes). Agents gather information by calling Windows APIs to collect data and
match it with the threshold values set in the Windows registry.
The Host Information Agent collects file system, system processor utilization, and running
program information from the system.
The Threshold Support Agent allows a management application, such as HP Insight
Manager, to set user-specified thresholds and alarms for monitored components (for
example, system processor utilization or file system utilization). This agent periodically
checks each monitored component and sends an alarm or trap to a management console,
when the number of errors for the component crosses the threshold defined in the
management application, .
The Version Control Information Agent collects information about the versions of HP support
software installed on the server.
For machines that are members of a cluster, the Clustering Information Agent collects
information about the cluster.
Based on the system configuration, resources, and component attributes, another sub-agent
calculates an optimal value for the storage space required to hold a memory dump when a
system crash occurs. The result of this calculation is compared to the current dump allocation
settings to determine and report if the allocated space is sufficient to hold the memory dump.
Output
• SNMP trap sent
• Windows System Log message (source is “Foundation Agent”)
• SMH shows status on performance page
Internal Operations Monitoring
Each agent tracks internal operation status by entering records in individual log files—for
example, evtagt.log, evtler.log, evtsync.log, FoundAgt.log, serverAgt.log—in the following
directory:
%windir%\system32\cpqmgmt
For example, the EvtSync service writes a record to evtsync.log when it cannot read the IPMI
System Event Log.
Note: The system typically creates these internal event logs when an error occurs and the
first record is written. If a particular kind of error does not occur, the log file for that
agent or service may not exist.
Page 12