You can configure the SystemEDGE agent to monitor specific Management Information Base (MIB) objects for thresholds or other exceptions. When you manage a large enterprise with hundreds of systems, you can configure the information that the agent monitors. The ability to configure the agent monitored information provides the management by exception that is necessary to maintain large-scale implementations without impacting management overhead.
Note: For information about collecting and monitoring capabilities of each agent, see Concepts chapter.
The maximum resolution of the monitor supported is 64 bit. The SystemEDGE agent provides the following types of monitoring:
Provides monitoring of any integer-based MIB object that the agent supports. You can create entries in the Self Monitor table to specify objects to monitor, comparison operators, threshold values, and severities, and the agent automatically monitors the objects according to the created entries. The agent monitors the objects on a polling interval, maintains a current state according to specified threshold and severity values. If the state change threshold exceeds, the agent sends a state change trap.
Note: For more information about self monitoring, see Self Monitoring chapter.
Provides monitoring of any single process instance or Windows service. You can create entries in the Process Monitor table to monitor whether a process or service is running or to monitor process table MIB objects against specified thresholds. The agent monitors the processes, maintains a current state according to specified threshold and severity values. If the state change threshold exceeds or the state of a process (running or stopped) changes, the agent sends a state change trap.
Note: For more information about process and service monitoring, see Process and Service Monitoring chapter. To monitor more than a single process instance, use a process group monitor because a process monitor can only monitor a single process instance.
Provides the ability to define a group of processes and monitor that group. You can create entries in the Process Group Monitor table defining process groups. The agent monitors the groups and sends a trap when the process group change. The agent monitors also provide process group metrics such as memory or cpu time consumed by all processes in a group. If necessary, a self monitor can monitor the process group metrics.
Note: For more information about process group monitoring, see Process Group Monitoring chapter.
Provides monitoring of any UTF-8 encoded system or application log file by searching for strings specified as regular expressions. You can create entries in the Log Monitor table, and the agent monitors the specified log file for lines matching user-defined regular expressions and sends a trap when a match occurs. You can associate a severity with the monitor, which is included with the sent trap. If you configure entries in the Log Monitor table accordingly, you can also monitor the size and number of files in a directory. You can monitor log file of size up to 2 GB.
Note: For more information about log file and directory monitoring, see Log File Monitoring chapter.
Provides monitoring of Windows event log entries using different filters, such as event source. You can create entries in the NT Event Monitor table, and the agent monitors the event log for events matching user-defined regular expressions and sends a trap when a match occurs.
Note: For more information about Windows event monitoring, see Windows Event Monitoring chapter.
Provides historical data collection for manager-side baselining and trend analysis. You can create entries in the History Control table, and the agent collects metrics over a time interval. The metrics collected is used to provide a picture of average system performance during a specific time interval.
Note: For more information about history collection, see History Collection chapter.
You can specify a threshold for the monitors to monitor 64-bit metrics. For example, if you specify a 64-bit threshold, such as 9,000,000,000,000,000,001 Bytes (9 * 10^18 plus 1) is a valid 64-bit value. However, you can define a threshold of 9 EB (Exabyte) by defining a scaling factor of 10^18 and a threshold of 9. The scaling factor defines the resolution of the monitor. You can define a scaling factor of 10^16, for a resolution of one percent, the corresponding threshold is 900.
If the scaling factor of a monitor is 2^32, for a 64-bit (2^64) metric, the maximum monitor value is 2^32.
SystemEDGE provides the process CPU utilization as a percentage of new variable, processTimePermil in the processTable. The Process Monitoring and Process Group Monitoring capabilities are provided using processTimePermil.
processTimePermil is the CPU permil utilization of a process over the last sample interval. The variable value is system-dependent. The value is the difference in the number of CPU ticks divided by the total CPU ticks. As SNMP does not support fractions, the value reported is multiplied by 1000. For example, a CPU permil of 505.833 is returned as 505833. To convert to percent, divide the returned value by 10000.
A new Aggregate table in the Systems Management MIB allows for aggregation of objects that have multiple monitors defined with different severities. The Aggregate table uses the new class, instance, and attribute properties in the Self Monitor and Process Monitor tables to determine whether monitors refer to the same object. The overall object state is determined by sending a state change trap for the breached monitor with the highest severity.
To relate an object instance to its hosted hierarchy, // and / are used as delimiters such as //hierarchy/instance. The hierarchy can be multi-tier such that object instances are supported for more than two tiers.
The hierarchy can be a host name //host/instance, it can be also of the form //<any1>/<any2>/.../<anyN>/<instance>. For object instances on local system, only instance can be specified and //./ is prepended automatically. You can prepend a namespace to identify virtual object instances such as 'lpar:'. For example, the object instance on an LPAR is represented as lpar://hmc1/hyper1/lpar1/disk1.
The super aggregation of object instances is provided only for the lowest two tiers and not all tiers. For example, //host/* aggregates all instances on a host. The special super aggregate //* is created (which aggregates all object instances) is dependent on the selected level of super aggregation.
|
Copyright © 2013 CA.
All rights reserved.
|
|