A First Example: Abnormal Dips in User Trafficedit

The easiest time series models have a unique feature. Past values are then used as a proxy to forecast future values. This unique feature is an io ie, input and output at the same time.

The following user_traffic.json file (Or .yml file if YAML format is used) is using:

  • One minute aggregations, ie bucket_interval equals 1m
  • A forecast horizon equal to 5 bucket_interval ie 5 minutes
  • A proxy, the span equal to 10 bucket_interval ie 10 minutes
  • A low anomaly type, since we want to detect dips in user traffic
{
  "bucket_interval": "1m",
  "default_datasource": "my-datasource",
  "features": [
    {
      "default": 0,
      "metric": "count",
      "field": "requests",
      "measurement": "traffic",
      "name": "count_all_requests",
      "anomaly_type": "low"
    }
  ],
  "interval": 60,
  "max_evals": 10,
  "name": "traffic-model",
  "offset": 30,
  "forecast": 5,
  "span": 20,
  "max_threshold": 90,
  "min_threshold": 50,
  "type": "timeseries"
}