Creates Alert Rules
Documentation Index
Fetch the complete documentation index at: https://handbook.fiddler.ai/llms.txt
Use this file to discover all available pages before exploring further.
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Unique identifier of the model.
Name of the Alert Rule.
The unique identifier of the metric for which we want to create an Alert Rule.
HIGH, MEDIUM, LOW The Supported bin sizes for the Alert Rules.
Hour, Day, Week, Month This API provides two alert evaluation methods, absolute comparisons determine if the current metric value breaches a predefined threshold, while relative comparisons analyze changes in metric by referencing a chosen time period.
raw_value, time_period The comparison condition while evaluating the Alert Rule.
lesser, greater The type of threshold algorithm used to evaluate alerts.
manual, standard_deviation_auto_threshold This field specifies the lookback period for relative comparisons, expressed as a multiple of the chosen time bin size. It determines the historical data timeframe used to evaluate changes in the metric value.
It indroduces the delay in the evaluation of Alert Rule. It is in multiple of hours, and max could be one year.
Unique identifier of the segment.
List of feature names for which we want to create an Alert Rule.
Setting Frequency Alerts on categorical columns requires specifying a category. For example, if the column represents geographical locations such as France, Germany, and India, you would pass 'France' as the category and 'geography' as the feature name to set a frequency alert.
It should only be specified for the drift Alert Rules. Is is unique identifier of the baseline selected for drift calculations.
Auto threshold parameters for the Alert Rule. These parameters are used to tweak the calculations of the thresholds.
Threshold value for triggering a warning alert.
Threshold value for triggering a critical alert.