Skip to main content

Glossary

The words Measured Cloud uses, in the order you meet them. Where the API or database uses a different machine name, it's given in brackets so you can match the docs to a payload.

The core workflow​

Data source​

One place your data lives — a Postgres database, a Snowflake account, an S3 prefix. Holds encrypted connection details and an execution mode: direct (Measured Cloud connects) or agent (the Privacy Agent connects from inside your network). (API: datasource.) See Data sources.

Data asset​

One table, view, query, or parquet dataset inside a data source. The home of its profile and its rules. (API: data-assets, table DataAsset.) See Data assets.

Rule​

One thing you verify about a data asset — "email is never blank", "row count is between 1 and 10 million". Every rule belongs to exactly one asset and carries a severity. Each rule is one Great Expectations expectation (rule_type + kwargs). (API: .../data-assets/{id}/rules.) See Rules.

Severity​

How much a rule's failure matters: Info, Warning, or Critical (the default). Alert rules route on it. See Rules → Severity.

Validation​

The runnable unit: a named set of rules, from one or more data assets, that run together, with a health score. (API: /api/v2/.../validations; some wire fields keep the older machine name rule_group.) See Validations.

Schedule​

A cron cadence attached to a validation, read in the organization's schedule timezone. Each firing starts an execution run. See Schedules.

Schedule timezone​

The one IANA timezone every schedule in an organization runs in, and the zone the app shows schedule times in. An org admin sets it in Settings → Scheduling (default UTC). See Set the schedule timezone.

Execution run​

One run of a validation — started by hand, by a schedule, or through the API. An asset's Run (trigger asset) and a single rule's Run (trigger rule) are execution runs of the asset's default validation. Fills in asset by asset and finishes with a status and a health score. A validation's runs are listed in its Run history, and each opens as a Validation run page. See Execution runs.

Health score​

The share of rule checks that passed: 100 × passed ÷ evaluated checks. For a validation or a data asset it covers every run in the last 30 days; on the dashboard it covers the selected period; on a single execution run it covers just that run. Its complement is the failure rate ("12% of checks failed in the last 30 days"). Not weighted by severity and not blended across runs. Good above 90, Fair 70–90, Poor below 70. See Health score.

Result​

The recorded outcome of a run: pass/fail per rule, failing counts, and sample failing rows. A validation run produces one validation-level result and one per asset. (API: validation-results, table ValidationResult.) See Results.

Alert rule​

"When this happens, tell these people": a scope (specific validations, all validations, or the whole workspace), a trigger (on_failure, on_success, always, on_anomaly), the severities it fires on, and a channel (Slack, email, Teams, PagerDuty, Jira, webhooks, and more). See Alerts.

Alert history​

Every firing of an alert rule — when it fired, the channel, whether it was Sent, Failed, or Pending, and any delivery error. Shown on each rule's page and across the workspace on the Alerts page. (Table AlertEvent.)

Around the workflow​

Profile run​

A statistical snapshot of a data asset — row count, schema, per-column null rates, distinct counts, ranges. Feeds column pickers, anomaly detection, and AI rule suggestions. See Profile runs.

Anomaly​

A statistical deviation in a tracked metric (row count, null percentage, schema) against a rolling baseline of profile runs — how Measured Cloud catches problems no rule was written for. See Anomalies.

Rule template​

A reusable set of rules not tied to any asset. Apply it to many assets, mapping columns where names differ. See Rule templates.

Custom SQL rule​

A rule written as a SELECT that returns the rows breaking it; an empty result passes. (Rule type unexpected_rows_expectation.) See Rules → Custom SQL rules.

Custom expectation​

A Great Expectations Expectation subclass from a Python package an operator lists in EXPECTATIONS_PLUGIN_PACKAGES. Appears in the rule catalog. See Rules → Custom expectation plugins.

Incident​

An internal ticket for a data-quality failure — assign, acknowledge, resolve, with a timeline. See Incidents.

Privacy Agent​

A process you run inside your own network that registers data sources with Measured Cloud and runs validation and profiling work there, so connection credentials never leave your environment. Authenticates with a dqa_… agent token. (Also called the DQ Agent in older material.) See Agent mode.

Recycle bin​

Where deleted data sources and data assets wait before they're purged. Restorable for RECYCLE_BIN_RETENTION_DAYS (30 by default). Validations and results are never binned. See Recycle bin.

Accounts and access​

Organisation​

The top-level tenancy boundary. Every organisation's data lives in its own database schema. See Multi-tenancy.

Workspace​

A grouping of data sources, assets, rules, validations, and members inside an organisation — for example one per department.

Team​

An optional grouping inside a workspace that scopes who sees which validations, schedules, and results.

Personal API key​

A dqk_… bearer token you mint under Settings → API keys for scripts, the SDKs, and MCP clients. See Create an API token.

Agent token​

A dqa_… token used by a Privacy Agent. Scoped to one organisation and optionally one workspace. Revocable.

Identity provider​

A per-organisation SAML or OIDC configuration for single sign-on: enabled, enforced (password login off), or disabled. See Configure SSO.

Invite​

A dqi_… one-time credential that adds someone to an organisation or workspace when accepted. Expires after INVITE_EXPIRY_DAYS.

Audit event​

A record of a security-relevant action — sign-ins, token mints and revocations, and writes and deletes to data sources, assets, rules, alert rules, and settings. Append-only: the database blocks edits, and deletes are allowed only for the dated retention prune. Queryable via GET /api/v1/organizations/{org_id}/audit-events.

LLM config​

An organisation's bring-your-own-LLM settings: provider, model, and an encrypted API key. See AI features.

Retired terms​

Rule suite​

Retired. Rules used to be grouped into named suites; they now live directly on their data asset, and a validation holds rules one by one instead of a list of assets. The API keeps rule-suites as a read-only alias for one release (its writes return 410 Gone), and fields such as suite_id stay in responses as null. See Rule suites (retired).

Checkpoint​

Retired. The runnable unit before validations. An asset's Run now runs its default validation, and each checkpoint maps to a validation that keeps its run history. The /checkpoints API is a deprecated read-and-run alias for one release (writes return 410 Gone). See Checkpoints (retired).