SelectZero Data Observability

SelectZero data testing

SelectZero runs SQL test cases for nulls, uniqueness, accepted values, ranges and custom queries, and compares results across sources, from sums and counts to full datasets row by row.

Compare: data testing tools · data comparison tools

• SQL Table Tests: All test cases are SQL-based validations with predefined types covering completeness (null checks), uniqueness, validity (accepted values, format), accuracy (value ranges), plus custom SQL query-based tests (docs).

• Table Comparison Tests: Comparison Tests allow comparing data from different data sources by comparing SQL results — numeric values (SUMs, COUNTs, AVGs), strings, and data sets row by row/column by column (docs).

• DataFrame Tests: SelectZero is a SQL-based platform that connects to data warehouses and databases; no in-memory DataFrame (Pandas/Spark) validation is documented (docs).

• Unstructured Data Tests: No unstructured or LLM-based column validation is documented (docs).

• Data Diff: No dedicated row-by-row data diff feature (comparing dev vs prod to surface inserted/deleted/changed records) is documented (docs).

• Data Contracts: No data contracts feature (machine-readable schema and quality SLA agreements between producers and consumers) is documented (docs).

• Inspect Failed Rows: Query + Total Tests return a count of invalid rows and total rows to calculate a pass rate; the docs do not document the ability to view actual failing rows (docs).

SelectZero data observability

SelectZero profiles columns and validates values, but it doesn't use ML anomaly detection. Checks are configured in the UI.

Compare: data profiling tools · data observability tools

• Data Profiling: Profiling computes column-level statistics including average, count of NULLs, count of NOT NULLs, unique values, and sum of values (docs).

• Field Monitors: Profiling tracks column-level statistics (null counts, unique values, averages) over time and flags deviations from historical patterns (docs).

• Distribution Monitors: Profiling analyzes value distributions over time — 'finding counts for each value' — and detects anomalies in how column values change across executions (docs).

• Custom SQL Monitors: Query-Based Tests allow defining an arbitrary SQL expression as a tracked validation with numerical or string comparison thresholds (docs).

• UI-Based Monitors: SelectZero is a UI-first platform; all checks and monitors are configured through the graphical interface without writing code (docs).

• Freshness Monitors: No dedicated freshness monitor (checking table last-updated timestamp against an expected ingestion window) is documented (docs).

• Volume Monitors: No dedicated volume monitor tracking row count changes over time is documented (docs).

• Schema Monitors: No schema change monitor detecting column additions, removals, or type changes is documented (docs).

• Segment Monitors: No segment/group-by monitor breaking down metrics by a dimension column is documented (docs).

• Partition-Based Monitors: No partition-based or time-filter scoping mechanism to limit scans to a recent slice of a partitioned table is documented (docs).

• Window-Based Monitors: No window-based monitor evaluating metrics over a configurable time window is documented (docs).

• Monitors-as-Code: No monitors-as-code capability (YAML/Python definitions version-controlled and deployed via CI/CD) is documented (docs).

• Pipeline Monitoring: No pipeline monitoring tracking orchestrator-level job statistics (dbt run times, Airflow DAG durations) is documented (docs).

• Anomaly Detection: The profiling page states users can 'detect deviations/anomalies in data and alert users', but no ML-based or threshold-based anomaly detection with dedicated monitors is documented — this appears to describe basic profiling deviation flags rather than a dedicated anomaly detection capability (docs).

SelectZero incident management

SelectZero sends alerts when tests fail. It has no incident workflow.

• Alert Notifications: Email-based alerts documented for test case execution and change events. Webhook notifications also supported for test suite failures per test-suites.html (docs).

• Incident Triage: No structured incident triage workflow (classify, assign owner, set priority, group related failures) is documented (docs).

• Alert Feedback: No alert feedback mechanism (marking anomalies as false positives, snoozing alerts) is documented (docs).

• Incident Resolution: No incident resolution lifecycle (root-cause context, mark as resolved, fix logging) is documented (docs).

• Ticketing Integrations: No Jira, ServiceNow, or Linear ticketing integration is documented. Alerting is limited to email and webhooks (docs).

• On-Call Integrations: No PagerDuty, Opsgenie, or similar on-call integration is documented (docs).

SelectZero data quality analytics

SelectZero shows test results on dashboards.

• Data Quality Dashboards: The Dashboard provides an overview of current data quality with status-based charts for test suites, business rules, objects, schedules, and alarms (docs).

• Data Health Scores: The dashboard shows status-based charts and alarm counts but no per-table or per-dataset quality health score (pass % metric) is documented (docs).

• Incident Resolution Metrics: No MTTD/MTTR dashboards or resolution-time/owner analytics are described anywhere on the site or pricing page; the platform focuses on validation alerts, not incident-response analytics (pricing).

SelectZero data lineage

SelectZero derives table- and column-level lineage automatically from queries, imports it from files, or lets you add it by hand.

Compare: data lineage tools

• Table-Level Lineage: Object-level lineage tracks relationships between catalog objects and can be visualized from the Catalog objects view (docs).

• Column-Level Lineage: Column-level lineage tracks field-to-field dependencies and can be opened from any column in the object-level lineage graph (docs).

• End-to-End Lineage: Lineage covers catalog objects and dbt-derived relations but does not span across the full data stack (ingestion tools like Fivetran/Airbyte through to BI tools like Tableau/Looker) (docs).

• Impact Analysis: No impact analysis feature showing which downstream assets are affected by an issue or code change is documented (docs).

SelectZero data governance

SelectZero's catalog syncs objects and columns, with a business glossary on paid plans.

Compare: data catalog tools

• Data Catalog: Built-in catalog with objects, columns, business glossary, and metadata management documented (docs).

• Data Catalog Integration: No integration with external catalog tools (Atlan, Collibra, Alation, Databricks Unity Catalog) is documented (docs).

• Data Classification: No automated PII/PHI/PCI data classification scanning column values is documented (docs).

• Data Products: SelectZero organizes work around a data catalog, lineage, and business glossary rather than grouping monitors/tests under a data-mesh-style domain or data product concept (docs).

SelectZero workflows

SelectZero has a REST API, webhooks, an MCP server and a dbt integration. It has no Python SDK or CI/CD checks.

• Scheduling: Full scheduling documented: selected days, monthly, custom interval (every N minutes), and CRON (UNIX 5-field and Quartz 6/7-field) (docs).

• Webhooks: Test suites support configuring one or more webhooks to notify external endpoints when an alert is triggered (docs).

• REST / GraphQL API: REST API documented with endpoints for test suites, test cases, profiling, connections, executions, users, and import — authenticated via API keys (docs).

• MCP Server: MCP server documented exposing catalog, lineage, glossary, and test case data via Streamable HTTP at /api/v1/mcp — compatible with Claude Code, Claude Desktop, VS Code Copilot, Cursor, and Microsoft 365 Copilot (docs).

• dbt Integration: dbt-derived catalog data and lineage are documented; the catalog supports dbt objects and their dbt-derived lineage with badge indicators for dbt relations (docs).

• CI/CD Integrations: No CI/CD pipeline integration (GitHub Actions, GitLab CI) blocking merges on check failures is documented (docs).

• Python SDK: No Python SDK is documented; programmatic access is via REST API only (docs).

SelectZero AI agents

SelectZero's AI assistant writes test case SQL and expected results from plain-language prompts. An MCP server is also available.

• AI Chat Assistant: AI Assistant documented for generating test case SQL and expected results from natural language prompts, and bulk-generating business descriptions for catalog objects and columns (docs).

• Natural Language Test Generation: SelectZero's AI assistant (launched in the 2024.8 release) lets users describe a validation rule in natural language and automatically generates the executable SQL validation, usable with any connected LLM (OpenAI, Copilot, Claude, etc.) (website).

• Automated Test Generation: Test generation is a user-directed manual workflow requiring explicit filter selection and user-triggered generation — not autonomous ML/statistics-based monitor creation without user prompts (docs).

• AI Issue Triaging: No AI-powered issue classification or root-cause analysis is documented (docs).

• AI Issue Resolution: No AI-driven remediation or fix proposal features are documented (docs).