Hi 🙂

ignore_small_changes

doesn’t work for column_anomalies, after taking a look in the source code, I don’t see this param among the parameters for this kind of anomaly detection. I want to track a specific column values, then alert if there is a drop or spike above 30%. I’ve been thinking that maybe add this config under

config

block, but I’m not sure. what are yours suggestions?

- name: stg_elementary_volumes
    columns:
      - name: row_count
        tests:
          - elementary.column_anomalies:
              column_anomalies:
                - sum
              timestamp_column: ts
              anomaly_sensitivity: 3
              anomaly_direction: both
              time_bucket:
                period: hour
                count: 1
              seasonality: hour_of_week
              tags: ['elem']
              config: 
                severity: warn
                ignore_small_changes:
                  spike_failure_percent_threshold: 30
                  drop_failure_percent_threshold: 30