Qualytics Data Observability

Qualytics data testing

Qualytics has more than 45 check types on SQL tables, from not-null to referential integrity and custom SQL. It compares tables, diffs field values between datastores, and shows the rows that fail.

Compare: data testing tools · data comparison tools

• SQL Table Tests: 45+ authored check types include Is Not Null (completeness), Is Unique, Between (accuracy), Matches Pattern (validity), ExistsIn for referential integrity (consistency), freshness checks, and custom SQL via SatisfiesExpression (docs).

• Table Comparison Tests: Aggregation Comparison check type compares aggregate metrics between two datastores or tables (docs).

• Data Diff: The Data Diff check type compares field values between two datastores to detect discrepancies (docs).

• Inspect Failed Rows: Source Records page shows failing rows with configurable sample size (10–10,000 records) for each anomaly (docs).

• DataFrame Tests: Qualytics operates as a SaaS/self-hosted platform that connects to datastores; there is no Python DataFrame (Pandas/Spark) testing library (docs).

• Unstructured Data Tests: No LLM-based semantic validation for unstructured text fields found in the check library or documentation (docs).

• Data Contracts: No machine-readable data contract feature (producer-consumer schema and quality SLA agreements) found in documentation (docs).

Qualytics data observability

Qualytics profiles data on a schedule, sets baselines and generates checks from them. Scans flag records that break checks as anomalies. Monitors are set up in the UI or as YAML config.

Compare: data profiling tools · data observability tools

• Data Profiling: Profile operations automatically compute field statistics and patterns, generating baseline metrics and AI Managed Checks (docs).

• Anomaly Detection: Automated anomaly detection during scan operations flags records that violate quality rules and surfaces them as actionable anomalies (docs).

• Freshness Monitors: Freshness tracking monitors when new data was last added or updated, with configurable Maximum Age thresholds and alert triggers (docs).

• Volume Monitors: Volumetric checks automatically track row counts and detect irregular changes against min/max thresholds derived from historical patterns (docs).

• Schema Monitors: Expected Schema check type asserts that a container conforms to a defined schema, detecting column additions, removals, or type changes (docs).

• Field Monitors: Metric Checks track changes in specific fields over time against min/max thresholds with temporal grouping (day, week, month) (docs).

• Custom SQL Monitors: Satisfies Expression check type executes custom Spark SQL expressions for arbitrary business rule validation (docs).

• Partition-Based Monitors: Max Partition Size and Min Partition Size check types assert record count limits per file or table partition (docs).

• Window-Based Monitors: The Time Distribution Size Check evaluates record counts within configurable time intervals (hourly, daily, monthly) and asserts that each window's count stays within defined bounds — enabling time-windowed data quality evaluation (docs).

• UI-Based Monitors: Full web UI for creating, editing, and managing checks, observability monitors, and anomaly triage (docs).

• Monitors-as-Code: Config-as-code exports the entire Qualytics configuration to a hierarchical YAML structure for version control and environment promotion (docs).

• Distribution Monitors: No statistical distribution monitoring (value distribution shifts, histogram comparison) found; Metric Checks use absolute/percentage thresholds only (docs).

• Pipeline Monitoring: No pipeline job monitoring features documented (docs).

• Segment Monitors: No segment-based monitoring (per-dimension or filter-based anomaly detection) explicitly documented (docs).

Qualytics incident management

Qualytics has a full incident workflow: alerts, triage, feedback, resolution, ticketing and on-call integrations.

• Alert Notifications: Flows system delivers real-time alerts via Slack, Microsoft Teams, Email, PagerDuty, and in-app notifications (docs).

• Incident Triage: Anomalies can be triaged by marking them as Acknowledged or Archived, with CLI automation for bulk triage workflows (docs).

• Alert Feedback: Anomalies can be marked as Invalid to flag false positives, providing feedback to tune the monitoring system (docs).

• Incident Resolution: Anomalies can be resolved and tracked through the full lifecycle; Agent Q provides guided remediation recommendations (docs).

• Ticketing Integrations: Native integrations with Jira and ServiceNow for creating tickets from anomalies via Flows (docs).

• On-Call Integrations: PagerDuty integration creates incidents with configurable severity when data quality events occur (docs).

Qualytics data quality analytics

Qualytics' dashboards and health scores track data quality.

• Data Quality Dashboards: Platform provides dashboards with quality scores, anomaly trends, observability metrics, and data health heatmaps (docs).

• Data Health Scores: Data Quality Score is computed per datastore and container, reflecting the percentage of passing quality checks (docs).

• Incident Resolution Metrics: The Insights dashboard tracks Quality Score dimensions, data volume, anomalies identified, and records/fields profiled/scanned — it does not include MTTD, MTTR, resolution time, or resolution-rate-by-owner analytics (docs).

Qualytics data lineage

Qualytics shows table- and field-level lineage from manual edges, catalog imports and warehouse metadata, with impact analysis.

Compare: data lineage tools

• Table-Level Lineage: Container-level lineage comes from five sources: manual edges, data catalog imports (Atlan, Alation, Collibra, DataHub, Purview), source collection (Snowflake, BigQuery, Databricks, Oracle), OpenLineage events, and computed containers (docs).

• Column-Level Lineage: Field-level lineage connects individual fields between containers and shows the active anomaly count per field; manual connections can link fields that source systems do not expose (docs).

• Impact Analysis: Container and datastore Lineage tabs offer an impact analysis of what a container feeds and what feeds it, ranked across a datastore and exportable as CSV (docs).

• End-to-End Lineage: Lineage spans multiple datastores (bronze/silver/gold tiers) and imports relationships from catalog integrations (Atlan, Alation, Collibra, DataHub, Purview), but does not directly ingest lineage from BI tools or orchestration platforms without an intermediate catalog (docs).

Qualytics data governance

Qualytics classifies data and syncs results to catalogs such as Alation, Atlan, Collibra, DataHub and Purview.

• Data Catalog Integration: Integrates with Atlan, Alation, Microsoft Purview, Collibra, and DataHub to surface quality insights in external catalogs (docs).

• Data Classification: Qualytics includes built-in data quality checks for detecting and classifying sensitive data patterns such as email addresses and Social Security Numbers in columns (docs).

• Data Catalog: Qualytics has no built-in data catalog; it integrates with external catalogs like Atlan, Alation, Purview, Collibra, and DataHub (docs).

• Data Products: Qualytics has "Datastore Grouping" to categorize datastores by environment/team/domain/priority, but each datastore belongs to only one group and it's positioned as operational organization, not a data-mesh-style data-product concept with rolled-up consumer-facing quality status (docs).

Qualytics workflows

Qualytics has a REST API, webhooks, an MCP server and a CLI that imports dbt test suites as checks, with GitHub Actions examples for promoting config and running scans. It has no Python SDK.

• Scheduling: Operations (profile, scan, catalog sync) can be scheduled with cron-based configurations (docs).

• CI/CD Integrations: CLI with GitHub Actions examples for config promotion, dry-run validation, and scheduled scans across environments; hard quality-gate blocking not confirmed (docs).

• Webhooks: Flows support webhook actions that send HTTP callbacks (POST/PUT) to external endpoints on quality events, with Basic, Bearer, and Digest authentication; n8n workflow automation is also supported (docs).

• REST / GraphQL API: REST API available for all platform operations including datastores, checks, anomalies, and operations (docs).

• MCP Server: Qualytics CLI includes an MCP server (qualytics mcp serve) exposing 35 tools for AI clients like Claude Code, Cursor, and Windsurf (docs).

• dbt Integration: The Qualytics CLI imports a dbt test suite as quality checks, and dbt pipelines call Qualytics through its API (docs).

• Python SDK: The Qualytics CLI is a command-line tool installed via pip; it is not a Python SDK for programmatic library usage (docs).

Qualytics AI agents

Qualytics has Agent Q, a chat assistant on every page that turns plain language into checks. It also generates checks automatically and triages issues. An MCP server is also available.

• AI Chat Assistant: Agent Q is a built-in chat interface (accessible as a floating widget on every page) that turns natural language descriptions into quality checks — explicitly documented as 'natural language into action — creating quality checks' (docs).

• Natural Language Test Generation: AgentQ lets users make a plain-language request inside quality-check/check-template dialogs, and it proposes check configuration settings that the user reviews and applies before saving ("Review-Before-Save Proposals") (docs).

• Automated Test Generation: AI Managed Checks automatically generate quality rules based on data profiling, which Qualytics AI writes and maintains (docs).

• AI Issue Triaging: Agent Q analyzes anomalies, identifies root causes, assesses impact, and provides recommended remediation steps (docs).

• AI Issue Resolution: Agent Q provides analysis and guidance but does not autonomously execute fixes; users must implement remediation steps manually (docs).