Skip to main content

Manual thresholds

It is possible to set manual thresholds for metrics. These thresholds can be created via API

There are two modes for the thresholds:

  • CAPPED caps the machine-learning corridor with the provided bounds and can only tighten the existing machine-learning range. This can be useful to imput theoretical limits for a metric.
  • STATIC disables machine-learning baselines for the target and uses the provided bounds directly

Static thresholds can be set at the level of a single metric instance or for all the metrics of a certain type. For more details see here.

Eyer will continue to investigate correlations also for the metrics that have manual thresholds enabled, so deviations rised based on manual thresholds can be combined to other deviations to form alerts.