Datastore Quality Score Troubleshooting¶

DO Calculating quality scores

This page documents common quality score problems and the steps to resolve them. Each section describes a symptom, its cause, and the resolution.

Missing Scores¶

The Datastore Has No Quality Score¶

A datastore shows no quality score on the Datastores listing page or its Overview tab.

Cause: None of its containers has a score yet, or the datastore is an enrichment datastore. Datastore-level quality scores are offered only for source datastores.

Resolution: For a source datastore, run Sync, then Profile, then Scan on the tables you want to monitor. Enrichment datastores do not get a datastore quality score.

A Container Shows `-` Instead of a Score¶

A table or file displays - where its quality score should be.

Cause: A container receives a score only after it has a completed profile and a completed scan. Containers showing - are excluded from the datastore score, so they do not pull it down.

Resolution: Run a Profile and a Scan that include the container. See Which Containers Get a Score?.

Unexpected Values¶

The Datastore Score Is Much Lower Than Expected¶

The datastore score is far below the scores of the tables you monitor.

Cause: Tables that were profiled and scanned but have no checks receive very low scores, and every scored table counts in the datastore average. This usually follows a datastore-wide Profile and Scan during setup.

Resolution: Run operations only on the tables you intend to monitor, or use tag weights so those tables dominate the average. See the low datastore score example for every option.

The Score Does Not Improve After Fixing the Data¶

You fixed the source data, but the score stays the same.

Cause: Qualytics does not know the data is fixed until a scan confirms it, and old anomalies count until they leave the decay window. Another dimension, or a dimension with a high weight, may also be holding the score down.

Resolution:

Run a Scan so the anomalies are re-evaluated.
Open the dimension breakdown on the container detail page and look for the dimensions that still score low.
Review the dimension weights for a dimension that dominates the score.

Coverage Keeps Dropping¶

The Coverage dimension of a container goes down over time even though no checks were removed.

Cause: Coverage rewards checks that scans assert often and recently. Checks left out of your scans, for example when scans assert only selected checks, count less as time passes.

Resolution: Pair targeted scans with a regular scan of all active checks.

Settings and Tags¶

A Settings Change Did Not Change the Score¶

You changed the decay period or dimension weights, but the displayed score is the same.

Cause: Settings changes apply to new scores only. Already-recorded scores keep the settings that were in effect when they were measured.

Resolution: Wait for the next scan, profile, or other event that recalculates the score. See What Triggers a Recalculation?.

Tagging a Container Did Not Change Its Weight¶

You added a weighted tag to a container, but the datastore score did not shift.

Cause: Container weights are recalculated only when tags are added to or removed from the datastore, or when you use Move to, Merge, or Remove on a tag's page. A tag assigned directly to a container does not change its weight until one of those happens.

Resolution: Change the datastore's tags, or use one of those actions on the tag's page, to recalculate the weights. See Tag Weight Modifiers.

AI Explanations¶

The Sparkle Icon and AI Explanations Are Missing¶

The Quality Score panel opens without the AI summary, and no sparkle icon appears next to the score.

Cause: AgentQ is not configured for the deployment. Without it, the panel shows the score, dimensions, and static descriptions, but none of the AI-written content.

Resolution: Ask a user who can manage integrations to configure AgentQ. See When AgentQ is not configured.