~Baseline Settings~

Users can customize baseline settings for each monitor, to suit organizational requirements. Activities and monitored attributes (MAs) can have individual baseline settings, or they can inherit the settings of the nodes above them. When you want to define separate baseline settings at the activity/MA level, verify that the Inherit checkbox is cleared.

Baseline Split Order

By default, the system calculates one baseline for each activity. When the Split Order feature is used, different baselines are calculated, according to predefined static and/or contextual attributes. For example, the baseline for an activity can be split according to server ID, location, etc. When the minimum number of measurements for a predefined value of an attribute is exceeded, a baseline is created for that attribute value.

Users define which attributes are included in the split, and the order in which the split is calculated, using the relevant Baseline tab. Split orders can be defined at the monitor, activity, and monitored attribute (MA levels).

The order in which attributes appear in the Split Order list determine the order in which the split is calculated. For example, if Location appears before Server ID, baselines for locations are created first, and then those baselines are split according to server IDs.

Measurements to Split

This value is the minimum number of measurements that must be obtained for a given attribute before a baseline is split for this attribute. For example, if the Measurements to Split value is 100, and Location is a selected attribute, once there are over 100 measurements for a specific location, a separate baseline is created for that location.

In order to be specified in the baseline split, each entity within an attribute category needs to have measurements that exceed the Measurements to Split value. For example, if Boston and New York have more than 100 measurements each for the number of days specified in the Use Data field under History, but other locations have less than 100 measurements, the baseline is split according to Boston and New York. The measurements for the other locations will be added towards the Global baseline for that attribute.

Note:

Once a baseline is split, it stays split.

Baseline Recalculation

To ensure that baselines always reflect current End User experience, Aternity recalculates the baseline on a regular basis. The baseline is automatically recalculated based on the values entered in the Recalculation fields. These values can be specified at the monitor, activity, and monitored attribute levels. The recalculation values are:

  • Every n days: Recalculation is done after the specified number of days pass (regardless of any interim changes in the baseline).
  • Change: Recalculation is done when the amount of new measurements exceed the specified percentage (regardless of when the baseline was last recalculated).

The system automatically recalculates the baseline every time either the days or the percentage condition is met.

Baseline History

Baselines are for normal conditions. For accurate detection, only normal condition measurements should be used. The history values are:

  • Use data: This uses the last n days for the baseline.
  • Not before: This allows you to exclude certain measurements from the baseline calculation by setting the date after the date of a known issue or change in the application or infrastructure, etc.