Algorithms
Rapport de recherche Datadog : Bilan sur l'adoption de l'informatique sans serveur Rapport : Bilan sur l'adoption de l'informatique sans serveur

Algorithms

Cette page n'est pas encore disponible en français, sa traduction est en cours.
Si vous avez des questions ou des retours sur notre projet de traduction actuel, n'hésitez pas à nous contacter.

Anomalies

FunctionDescriptionExample
anomalies()Overlay a gray band on the metric showing the expected behavior of a series based on past.anomalies(METRIC_NAME>{*}, '<ALGORITHM>', <BOUNDS>)

The anomalies() function has two parameters:

  • ALGORITHM: Methodology used to detect anomalies.
  • BOUNDS: Width of the gray band. bounds can be interpreted as the standard deviations for your algorithm; a value of 2 or 3 should be large enough to include most “normal” points.

Note: If you are using the agile or robust anomaly detection algorithms with weekly or daily seasonality, you can update your anomaly detection monitor to account for a local timezone using both the API and the UI.

Here’s a two-minute video walkthrough:

See the Anomaly Monitor page for more info.

Outliers

FunctionDescriptionExample
outliers()Highlight outliers series.outliers(<METRIC_NAME>{*}, '<ALGORITHM>', <TOLERANCE>, <PERCENTAGE>)

The outliers() function has three parameters:

  • ALGORITHM: The outliers algorithm to use.
  • TOLERANCE: The tolerance of the outliers algorithm.
  • PERCENTAGE: The percentage of outlying points required to mark a series as an outlier (available only for MAD and scaledMAD algorithms)

See the Outlier Monitor page for more info.

Forecast

FunctionDescriptionExample
forecast()Predicts where a metric is heading in the future.forecast(<METRIC_NAME>{*}, '<ALGORITHM>', <DEVIATIONS>)

The forecast() function has two parameters:

  • ALGORITHM: The forecasting algorithm to use - select linear or seasonal. For more information about these algorithms, see the Forecast Algorithms section.
  • DEVIATIONS: The width of the range of forecasted values. A value of 1 or 2 should be large enough to forecast most “normal” points accurately.

A number of the graphing options disappear, as forecasts have a unique visualization. After successfully adding Forecast, your editor should show something like this:

Other functions