Try StarTree Cloud: 30-day free trial
Alert tuning for accuracy

How to tune alerts for improved accuracy using Anomaly Filters

Link to concepts - using "Anomaly filters" as part of post-processor

These filters can be defined during the alert creation. link to how to create alerts

Day of the week Filter:

Using this filter you can define the day of the week that will be used to filter anomalies to reduce false alarms. These days are usually high volume or low volume and you dont want to detect those anomalies unnecessarily wiki page: /docs/startree-enterprise-edition/startree-thirdeye/reference/operators/post-processor/time_of_week DaysOfWeek: ["MONDAY", "SATURDAY"], HoursOfDay: [0,1,2,23], DayHoursOfWeek: {"FRIDAY": [22, 23], "SATURDAY": [0, 1, 2]}

Offsets Filter:

Using this filter you can define an offset to ignore those anomalies during alert creation/configuration so that it will help reduce false alarams during this time Metric Name: "Business KPI", # customize label Offsets: ["P7D", "P14D"], # parameter of the detector Aggregation: "MEDIAN" # parameter of the detector

Cold Start Filter:

Usually model predictions are impacted by low data volume in the beginning. With this feature users can define from which day/time there is good for the model to start learning to predict anomalies accurately coldStartPeriod": "P14D", # filter anomalies in the first 14 days of the dataset tableName: "my_dataset"

Events Filters:

Using this filter, users can apply the event filters to filter anomalies on those specific holiday period to reduce false alarms BeforeHolidayMargin: "PT4H", AfterHolidayMargin: "PT4H"

Threshold filter:

Using this filter, users can apply a threshold on the record volume to reduce false alarms Min limit: "100", Max limit": "10000000", Metric Name: "controlMetric", # customize label Timestamp: "ts", # time column of the side input Threshold metric: "sideMetric" # column to threshold on in the side input

Time of the week filter:

Using this filter users can specify with day of the week and the time window during which the anomalies will be ignored to reduce false alarms Days Of Week": ["MONDAY", "SATURDAY"], Hours Of Day": [0,1,2,23], Day Hours Of Week": {"FRIDAY": [22, 23], "SATURDAY": [0, 1, 2]}