Managing resource monitors

A Resource Monitor is a set of thresholds used to evaluate a metric. The evaluation of a metric sample against the thresholds marks the sample as GOOD, WARN or POOR. Resource Monitors allow automatic evaluation of metric samples.

This topic contains the following sections:

BESTHIGHER and BESTLOWER directions of a metric

A Resource Monitor is associated with a specific direction, based on its value from a capacity-planning point of view. These directions are:

BESTHIGHER: If a higher value is better than a lower value. For example, free disk space.
BESTLOWER: If a lower value is better than a higher value. For example, CPU utilization.

According to the direction of the metric, a monitor specifies two thresholds for the resource:

For BESTLOWER metrics:
- The good threshold is the limit under which a metric value is GOOD; any value greater than this threshold (and less than the following) will be marked as WARN.
- The warn threshold sets the limit over which a metric value is to be considered POOR
For BESTHIGHER metrics:
- The good threshold is the limit over which a metric value is GOOD; any value lower than this threshold (and more than the following) will be marked as WARN.
- The warn threshold sets the limit under which a metric value is to be considered POOR.

Example

A monitor can state that the metric CPU_UTIL of type BESTLOWER is considered:

GOOD between 0% and 50%
WARN between 50% and 75%, and
POOR over 75%

While the metric FREE_DISK_SPACE of type BESTHIGHER is considered:

GOOD over 5GB
WARN between 1GB and 5GB, and
POOR under 1GB.

Tracking samples

Once the monitor is in place, KPI samples are classified by the warehousing process and additional statistics are produced in order to track the number of:

Good samples (Ng)
Warning samples (Nw)
Poor samples (Np)
Missing/invalid samples (Nm).

Sample percentages

Over a specified time range, the sum of the resource counters listed above is equal to Ns (the number of samples in the time range).

Therefore, you can define the poor sample percentage as:

Poor% = Np/Ns

Similarly, you can define the the warning and good sample percentages as:

Warning% = Nw/Ns and Good% = Ng/Ns

The comparison of these percentages over a period of time is called threshold violation analysis.

Example: A threshold violation analysis, reporting the CPU utilization % of a server in the last week

In the example, for 18% of the time the CPU Utilization has a poor quality level, according to the monitor definition.

Note

Data samples are classified by the warehousing process. A change in threshold values is applied to new samples only.

Monitor levels

Monitors can be defined at two levels:

Generic monitors: Valid for all systems and business drivers. They can only be defined by the BMC TrueSight Capacity Optimization administrator. For more information, see Configuring-alerts-and-recommendations.
Specific monitors: Valid for a specific set of systems or business drivers. If a specific monitor exists, it overrides the generic monitor for the same system or business driver, allowing a more granular control on the thresholds applied by the monitor.

Viewing existing resource monitors

To view a list of specific monitors active on an entity, click Resource Monitor in the entity's main page.

Note

A wildcard (*) is used to indicate that the monitor applies to any value for the column in which the symbol appears.

Adding and configuring a new resource monitor

To add a new monitor, click Add monitor in the entity's main page.

To configure the new monitor, follow these steps:

Select the resource metric to which the monitor should be applied.
Select the subresource (optional).
Specify values for the GOOD and WARN thresholds.
Specify the validity of the monitor (default: always).

Where to go from here

Adding and configuring Resource Monitor rules to provide recommendations

Examples-of-formulas-and-parameters-in-custom-Optimizer-rules

Viewing-alert-log-messages