2.2.3 Threshold monitoring

(1) Threshold monitoring methods

The following two threshold monitoring methods are available:

• Fixed threshold method

  This method sets and uses a fixed threshold. Use this method for normal system monitoring operation.

  Note that determining a fixed threshold at times such as the start of operation might be difficult. Therefore, for the resources that can be collected by SSO, default values are provided in the threshold definition file. However, these default values are simply reference values and must be optimized if needed.

• Statistical threshold method

  This method theoretically calculates threshold values from the standard deviation based on statistics. How much collected data indicating warning or critical is included is specified as the ratio of data, treating the actually collected data as the statistical target parameter. The collection data that is the statistics target is specified as statistical total time. Since the statistical target parameter changes as the collection progresses, specify the calculation timing so that the threshold also changes at a given timing. You can also calculate the threshold by time zone by dividing the statistical period into time zones. The ID given to each time zone is referred to as time zone ID.

For the fixed threshold method, you must consider and set appropriate threshold values. For the statistical threshold method, you must consider and set the ratio of appropriate data. Generally, users can more intuitively understand fixed thresholds than statistical thresholds. Therefore, the fixed threshold method is preferable for the design of a monitoring system.

Regardless of the method, to change the threshold according to the time zone, you must link a task scheduler (Windows), cron (Linux), or a product that has a scheduler function (JP1/AJS3) with the SSO operation commands.

To Page Top

(2) Thresholds and resource statuses

Threshold monitoring of the resource monitoring function can monitor whether the resource collection value exceeds the threshold. You can set a warning threshold and a critical threshold. With these thresholds, you can check the resource status (Normal, Warning, or Critical). Note that there are some resources whose threshold is often exceeded momentarily. In such a case (in the event of a singularity), you can also set a continuous over counter so as not to change the resource status.

The following figure shows the relationship between the thresholds (warning and critical) and the resource statuses. This figure is an example for when the continuous over counters of the warning threshold and the critical threshold are both set to 3.

Figure 2‒14: Relationship between threshold and resource status

The following table describes how the resource statuses are judged.

Table 2‒7: How resource statuses are judged

Status	How the status is judged
Normal	The resource value is below#1 the warning threshold value. Alternatively, the number of times the resource value has consecutively gone above#2 the warning threshold does not exceed the value of the warning-threshold continuous over counter.
Warning	The number of times the resource value has consecutively gone above#2 the warning threshold exceeds the value of the warning-threshold continuous over counter, but the conditions for judging the status as Critical have not been met.
Critical	The number of times the resource value has consecutively gone above#2 the critical threshold exceeds the value of the critical-threshold continuous over counter.
Unknown	The collection status has become Postponing.
Non-monitoring	The collection status has become Completed, Impossibility, or Standing by. The number of monitored instances being collected has decreased.

A resource status change event can be issued when the resource status changes. For details on the events, see G. Events.

To Page Top

(3) Threshold verification

During initial deployment and the initial stage of operation, it is difficult to determine the optimal fixed thresholds and the ratio of data. Therefore, you might have to revise the thresholds (fixed thresholds or ratio of data). At this time, you can use the actual collection data to check how many times the resource status has changed to Warning or Critical based on certain values specified as thresholds and continuous over counters. You can specify the current values or any values for the thresholds and continuous over counters. After verification, the specified values can also be used for collection conditions.

To Page Top

(4) Automated action

Actions can also be automatically executed when the resource status changes. An automated action is a function that automatically starts commands that execute operations such as notification to the system administrator and data collection. You can also define variables in these commands. For details on the variables you can define, see H. Variables That Can Be Defined via Automated Action. Commands can be executed by superuser in Linux or by Administrator in Windows.

To Page Top

(5) Notes

To Page Top