A severity level says how much a failing test matters. Each level has a weight, and the weight decides how much a test's runs count in the quality score. Each test has one severity level.
Manage the levels on Settings > Trust Scores > Severity Levels, where you can add, rename and delete levels, change their weights and see how many tests use each. Set a test's severity in the test's settings.
Default levels
| Severity | Weight | Meaning |
|---|---|---|
| Critical | 16 | Expected to pass. A failure is a critical business risk with significant downstream consequences and needs to be addressed immediately. |
| High | 9 | Expected to pass. A failure likely has several downstream impacts, and the data should not be used until the issue is resolved. |
| Medium | 4 | Expected to pass. A failure means the data is untrustworthy and should be used with caution. |
| Low | 1 | Expected to pass, but a failure does not mean significant downstream impact or definitively untrustworthy data. |
| Informational | 0 | Provides information. Failures are not necessarily unexpected, and the test is not meant to affect quality scores. |
The default weights are squares (1, 4, 9, 16), so each step up counts for much more than the one below: one Critical run outweighs 16 Low runs.
How weights are used
Quality Score = Σ (run score × severity weight) ÷ Σ (severity weight)
The weight multiplies every run of the test in the window, so a test's influence depends on its severity and on how often it runs. See How Trust Scores Are Calculated for the full calculation and a worked example.
- Informational (weight 0) runs add nothing to either side of the quality score. They still count toward the coverage score and toward Total Executions.
- Changing a weight applies to every run in the window at the next trust score calculation, and immediately to the KPI bar. The Test Execution Activity panel keeps the weight each run had when it started.
- Changing a test's severity applies to its runs from then on. Runs already made keep the severity they ran with.
Choosing severities
Severity is the main lever you have over the quality score, so agree the meaning of each level before tests are written. A useful check: if every test on a table is Medium, the weights cancel out and the quality score becomes a plain pass rate. Differences in severity are what make a Critical failure stand out.