Coalesce Data Quality

Coalesce data testing

Coalesce Quality checks SQL tables, compares tables, and reconciles them row by row with checksums and bisection to narrow down mismatches. You can inspect the rows that fail.

Compare: data comparison tools

• SQL Table Tests: SQL tests from templates (not_null, unique, accepted_values, relationships) plus custom business queries whose returned rows count as failures (docs).

• Table Comparison Tests: Reconciliation compares a source and target dataset by row count or aggregates, across databases, clouds or accounts (docs).

• Data Diff: Reconciliation in row_checksum mode (the default) uses checksums and bisection to narrow a mismatch down to the rows or groups that differ (docs).

• Inspect Failed Rows: Failed records from SQL tests are stored in an audit table for inspection (docs).

• DataFrame Tests: Checks run as SQL against connected warehouses; no in-memory DataFrame validation library is offered (docs).

• Unstructured Data Tests: No checks on free-text or document data are documented (docs).

• Data Contracts: No data contract feature is documented; ownership and expectations are managed through data products (docs).

Coalesce data observability

Self-learning anomaly detection adjusts to how volatile your data is, monitors are defined in the UI or in YAML, and it alerts on failed dbt, SQLMesh, Coalesce and Airflow runs.

Compare: data observability tools

• Anomaly Detection: Self-learning anomaly detection learns patterns and adjusts sensitivity to data volatility (docs).

• Freshness Monitors: Freshness monitors detect stale data (docs).

• Volume Monitors: Volume monitors detect unexpected row count changes (docs).

• Schema Monitors: Schema monitors detect structural changes (docs).

• Field Monitors: Field stats monitors track statistical properties of fields (nulls, distribution) and learn a time-series pattern for each metric (docs).

• Distribution Monitors: Field stats monitors detect unexpected changes in field distributions (docs).

• Segment Monitors: Monitors can be segmented to detect a drop in one segment of the data (docs).

• Custom SQL Monitors: Custom SQL monitors track business metrics with anomaly detection (docs).

• Partition-Based Monitors: Time partitioning limits each monitor run to a time slice of the table, with a Partition Timeline tool (docs).

• UI-Based Monitors: Monitors can be placed and configured in the UI (docs).

• Monitors-as-Code: Monitors can be defined in YAML and deployed with synqcli (docs).

• Pipeline Monitoring: Alerts on dbt, SQLMesh and Coalesce Transform model run failures and Airflow task failures (docs).

• Data Profiling: Field stats monitors track column statistics over time, but a point-in-time profiling view is not documented (docs).

• Window-Based Monitors: Time partitioning sets the window a monitor measures (e.g. yesterday's data), but tumbling or sliding windows are not documented (docs).

Coalesce incident management

Coalesce Quality triages incidents, takes feedback on alerts, and pages on-call teams. It has no ticketing integration.

• Alert Notifications: Alerts go to Slack, Microsoft Teams, email, PagerDuty, Opsgenie, owners and webhooks (docs).

• Incident Triage: Incidents can be declared across monitors, tests and pipeline failures, with owners, status and comments (docs).

• Alert Feedback: Monitor feedback on detected anomalies teaches the model business context and reduces false positives (docs).

• Incident Resolution: Incident management tracks resolution from start to finish, with downstream impact from lineage (docs).

• On-Call Integrations: Integrates with PagerDuty and Opsgenie (docs).

• Ticketing Integrations: No Jira, Linear or ServiceNow integration is documented (docs).

Coalesce data quality analytics

Coalesce Quality's dashboards, health scores and resolution metrics track data quality.

• Data Quality Dashboards: Analytics shows issue trends, quality dimensions and test coverage, filterable by owner, data product and platform (docs).

• Data Health Scores: Data products get a quality score, with scores per data quality dimension (docs).

• Incident Resolution Metrics: Analytics reports MTTR and new vs resolved issues over time (docs).

Coalesce data lineage

Coalesce Quality extracts column-level lineage from the warehouse and BI tools several times a day, with impact analysis.

Compare: data lineage tools

• Table-Level Lineage: Lineage is extracted automatically from the warehouse and BI tools several times a day (docs).

• Column-Level Lineage: Column-level lineage from BigQuery, Snowflake, Redshift and ClickHouse to Looker and Tableau (docs).

• End-to-End Lineage: Lineage spans the warehouse, transformation layers (dbt, SQLMesh, Coalesce Transform) and BI tools (docs).

• Impact Analysis: Shows the downstream impact of an issue or an upstream column change across transformations and BI (docs).

Coalesce data governance

Coalesce Quality has a built-in catalog that shows who owns each table and who uses it through which tool, and data products that give each team ownership of its data. It syncs with Atlan and Unity Catalog.

• Data Catalog: Built-in catalog with search, ownership and query activity showing who uses each table and through which tool (docs).

• Data Catalog Integration: Syncs with Atlan and Coalesce Catalog (docs).

• Data Products: Data products group business-critical assets with monitors, uptime and ownership (docs).

• Data Classification: No classification of sensitive data is documented; check categories classify checks, not data (docs).

Coalesce FinOps

• Cost Monitoring: No warehouse cost monitoring is documented (docs).

• Query Performance: Query activity shows who queries data and what builds each table, but query performance monitoring is not documented (docs).

Coalesce workflows

Coalesce Quality has a REST API, webhooks, an MCP server and a dbt integration. It connects to GitHub, GitLab, Bitbucket and Azure DevOps, and runs reconciliation in CI. It has no Python SDK.

• Scheduling: Scheduled monitors control when they run and which time window they evaluate (docs).

• CI/CD Integrations: Code integrations with GitHub, GitLab, Bitbucket and Azure DevOps, and reconciliation suites run in CI/CD (docs).

• Webhooks: Alerts can be sent to a webhook target (docs).

• REST / GraphQL API: Developer API covering monitors, alerts, SQL tests, lineage and entities (docs).

• MCP Server: Scout exposes an MCP server (docs).

• dbt Integration: Integrates with dbt Core and dbt Cloud, including test failures and model run failures (docs).

• Python SDK: Provides CLIs and an API, but no Python SDK (docs).

Coalesce AI agents

Coalesce Quality's Scout agent suggests tests, triages issues, and opens pull requests with code fixes and recommended tests. An MCP server is also available.

• Automated Test Generation: Deployment rules apply monitors automatically to matching assets, and Scout recommends and deploys tests based on lineage (docs).

• AI Issue Triaging: The Scout agent triages every open issue, with automated prioritization and noise filtering (docs).

• AI Issue Resolution: Scout generates ready-to-ship code suggestions and deploys recommended tests as pull requests; engineers approve (docs).

• AI Chat Assistant: Scout can be used from MCP clients, but a built-in chat assistant is not documented (docs).

• Natural Language Test Generation: Not documented (docs).