Anomaly Impact¶

Reviewing anomaly impact

Every anomaly carries an Impact: Low, Medium, High, or Critical. Impact ranks how much an anomaly matters, so you know which ones to review first. Anomaly lists are sorted by it, and Flows can start only for anomalies at the level you choose.

How Impact Is Decided¶

Impact comes from the checks that failed in the anomaly. For each failed check, Qualytics:

1. Rates how much failed.
2. Adjusts that rating by the check's Importance.

The anomaly then takes the highest level among its failed checks. Levels are compared, never added up, so an anomaly with many Low failures is still Low.

How Much Failed¶

Each failed check is rated by whichever of these gives the higher level:

Failed records: the share of the checked records that failed the check in the scan.

A check that tests the table as a whole rather than individual records, such as Expected Schema, Freshness, or Volumetric, is rated High when it fails. A check whose rule does not report how many records failed is rated Low.

Impact metrics: when the check has impact metrics with an expected range, a metric outside its range rates the failure High, and a metric far outside it rates it Critical.

Adjusted by Check Importance¶

The rating is then adjusted by the importance of the check that failed. High and Critical checks can raise it, and a Low-importance check can lower it; a Medium check leaves it as it is. A Critical check's failures are never Low, and a Low-importance check's failures are never Critical.

To make a check's anomalies rank higher or lower, change its importance through its tags. See Check Importance.

How Was This Decided?¶

Hover over an anomaly's impact to see an explanation of its level. Under How was this decided?, the explanation lists what set the level, for the failed check that decided it:

• Failed records: the share of the checked records that failed, or that a table-level check failed.
• Check importance: the check's importance, shown only when it raised or lowered the level. When several checks failed, it names the check.
• Impact metrics: the metric that is outside, or far outside, its expected range.

In the Failed Checks section of the anomaly, each failed check also shows its own impact, next to its impact metrics.

When Impact Updates¶

Impact is decided when the anomaly is created. While the anomaly is open (Active or Acknowledged), it is decided again whenever the importance of one of its failed checks changes because of tags (a tag is added to or removed from the check, a tag's weight is changed, or a tag is deleted) or because an AI Managed check is converted to Authored. The update runs in the background, so the new level can take a moment to appear.

Tags on the anomaly itself do not change its impact. Only the tags on the checks that failed do, through the checks' importance.

Where Impact Appears¶

• Anomaly details: the Impact value in the Summary section, and the impact of each failed check. See Anomaly Insights.
• Anomaly lists: the Impact sort, which is the default order, and the Impact filter. See Filter and Sort Anomalies.
• Flow triggers: the Anomaly Impact filter on the Anomaly Detected and Anomaly Status Changed triggers, which starts the Flow only for anomalies whose impact is at least, at most, or exactly the level you pick. See Configure a Trigger.
• Timeline: impact changes, credited to the person whose action caused them with a Qualytics co-author badge, or to Qualytics alone otherwise.
• API: the impact_level field of an anomaly, and the impact_level filter of GET /api/anomalies. See Anomalies API.

See Also¶