Skip to content

Getting Started with Anomalies

When a Scan finds data that fails a quality check, Qualytics records an anomaly. Triage is what you do next: find the anomaly, read its evidence, decide who owns it, and set a status that says where it stands. This page walks through a first triage from start to finish, so you know what each part of an anomaly is for before you open the detailed pages.

Show me how

The app can walk you through this. Press H on an open anomaly's page and pick Resolve an Anomaly: the walkthrough highlights each step while you do the real work. See The How-tos Scope for every way to start one.

Find the Anomaly

Anomalies are listed wherever the data they belong to is shown:

  • Explore › Anomalies lists the anomalies of every datastore you have access to.
  • A datastore's Anomalies tab lists the anomalies of that datastore.
  • A table's or file's Anomalies tab, and a field's, list only the anomalies found there.
  • A check's Anomalies tab lists the anomalies that check produced, and a check template's Anomalies tab lists those produced by every check built from the template.

Every list opens on its Open tab, sorted by Impact so the anomalies that matter most come first. You can also arrive from a notification: a Flow can send one when a Scan finds an anomaly, and you are notified when someone updates an anomaly you are assigned to. See How Anomalies Work.

Read It

Click an anomaly to open its details. The page is organized in sections, and each answers a question:

Section What it tells you
Summary What kind of anomaly it is, how many records it affects, how many checks failed, its Impact, when and by which Scan it was found, and where the data lives.
Description A plain-language explanation of the issue. Anyone with the Author team permission can rewrite it.
Failed Checks Each check that failed, with its failure message. Click one to see the check without leaving the page.
Source Records The rows behind the anomaly, when the datastore stores them.
Tickets Jira or ServiceNow tickets linked to the anomaly.
Timeline Every change made to the anomaly, and the team's comments.

An anomaly is evidence of a potential issue, not proof that the data is wrong. Read the failed check messages and the source records alongside what you know about the data before deciding what the anomaly means. See Anomaly Views for each section in detail.

Assign It

The Assignees in the Summary section name the people responsible for the anomaly. If the check that failed has a default assignee, that user is already there. Add yourself or a teammate when someone needs to take ownership. Assignees are notified each time someone else changes the anomaly. See Add Anomaly Assignees.

Set Its Status

A new anomaly starts as Active. Once you have looked at it, pick the status that says where it stands:

  • Acknowledge it when you have reviewed it but it still needs work. It stays in the open list, marked as seen. See Acknowledge Anomalies.
  • Archive it when the work is done, with the resolution that says why: Resolved for a real issue that was fixed, Invalid for one that was not a real issue, Duplicate for one already covered by another anomaly, or Discarded for one that is no longer relevant. Add a comment to explain the choice. See Archive Anomalies.

An archived anomaly can be restored later. It returns to the open list as Acknowledged. See Anomaly Status for every status and how a later Scan treats each one.

Deep Dive

Understand what an anomaly is, how a Scan creates it, how it is triaged, and what each part of it shows.

  • Introduction

    What an anomaly is, why it matters, and a short tour of every topic in this Deep Dive.

    Introduction

  • How It Works

    How a Scan creates anomalies, AI Managed and Authored checks, the lifecycle, duplicates, and notifications.

    How It Works

  • Record Anomalies

    Anomalies that flag one row: when they are created, how rows are grouped, and examples.

    Record Anomalies

  • Shape Anomalies

    Anomalies about many rows or the whole dataset: what creates them, what they count, and examples.

    Shape Anomalies

  • AI Managed and Authored

    How anomalies from AI Managed and Authored checks differ, the AI badge, and how triage tunes AI Managed checks.

    AI Managed and Authored

  • Impact

    How the Low to Critical impact level is decided from the failed checks, and when it updates.

    Impact

  • Status

    The lifecycle of an anomaly, from Active and Acknowledged to the archived statuses.

    Status

  • Assignees

    How the Assignees field works, who can be assigned, and when assignees are notified.

    Assignees

  • Fingerprints

    How Qualytics recognizes a newly detected anomaly as a duplicate of one it already found.

    Fingerprints

  • Anomaly Views

    The four screens that show anomalies: how to reach each one, what it shows, and what you can do there.

    Anomaly Views

  • Description

    The business-friendly explanation of the data quality issue behind an anomaly.

    Description

  • Source Records

    The rows behind an anomaly, how to read them, and how masked values are protected.

    Source Records

  • Tickets

    Linked Jira and ServiceNow tickets: sync modes, what reaches the ticket, and status sync.

    Tickets

  • Timeline

    The record of every change to an anomaly, and how comments and mentions work.

    Timeline

  • Examples

    Real scenarios that follow an anomaly from the Scan that finds it to the decision that closes it.

    Examples

  • Best Practices

    Guidelines for keeping the anomaly list short, trusted, and acted on.

    Best Practices

  • Permissions

    The team permission and user role behind every anomaly action.

    Permissions

How-tos

Triage and manage anomalies step by step, grouped by what each task changes.

Managing

  • Acknowledge Anomalies

    Mark an anomaly as reviewed while it stays in the open list.

    Acknowledge Anomalies

  • Archive Anomalies

    Close an anomaly as Resolved, Duplicate, Invalid, or Discarded.

    Archive Anomalies

  • Restore Anomalies

    Bring an archived anomaly back to the open list as Acknowledged.

    Restore Anomalies

  • Delete Anomalies

    Permanently remove archived anomalies that are no longer needed.

    Delete Anomalies

  • Edit Anomaly Description

    Rewrite the explanation shown in an anomaly's Description section.

    Edit Anomaly Description

Tags

  • Add Tags

    Apply tags to an anomaly to categorize it.

    Add Tags

  • Remove Tags

    Take tags off an anomaly.

    Remove Tags

  • Bulk-Edit Tags

    Add or replace the tags on many anomalies at once.

    Bulk-Edit Tags

Assignees

Comments

Tickets

Source Records

Reference

  • API

    Read, update, archive, and delete anomalies, and work with their comments, source records, and tickets programmatically.

    API

  • FAQ

    Common questions about types, Impact, status, duplicates, tags, tickets, and assignees.

    FAQ