Table Monitor

Broad anomaly detection for freshness, volume, and schema changes across many tables at once.

Overview

Get broad, low-config anomaly detection for freshness, volume, and schema changes across many tables at once. Point it at a database and schema, and Monte Carlo's ML handles the rest on its collection cycle. Table monitors use metadata instead of querying the table, keeping compute costs minimal.

📘
Reference scope
This page covers MaC YAML configuration. For how table monitors work and sensitivity settings, see Table Monitors.

Table monitors have no user-defined schedule or explicit thresholds. Alerting is anomaly detection only, controlled by sensitivity. For fixed thresholds or custom metrics on a single table, use a metric monitor.

Quick Start

montecarlo:
  table:
    - name: analytics_health
      description: Monitor all analytics tables for freshness and volume
      asset_selection:
        databases:
          - name: analytics
      alert_conditions:
        - metric: last_updated_on
        - metric: total_row_count
      domains:
        - my-domain

Create interactively with create_or_update_table_monitor(dry_run=True) via the Monte Carlo MCP server.

Configuration

description — what this monitor covers

string · required

Displayed in the Monte Carlo UI and in incident notifications. Max 512 characters.

description: Monitor all analytics tables for freshness and volume

asset_selection — which tables to monitor

object · required

Select tables by database, schema, name patterns, or tags. Combine databases with filters and exclusions for precise control.

asset_selection:
  databases:
    - name: analytics
      schemas: [core, staging]
  filters:
    - table_name: fct_
      table_name_operator: STARTS_WITH
  exclusions:
    - table_name: _deprecated
      table_name_operator: ENDS_WITH

Properties

databases — databases to monitor

array of objects · required

Each entry selects a database and optionally narrows to specific schemas.

Property	Type	Required	Description
`name`	string	yes	Database name
`schemas`	array of strings	no	Schemas to include. Omit to include all.
`tables`	object	no	Not supported — the backend rejects table-level selection for table monitors. Use `filters` instead.

databases:
  - name: analytics
    schemas: [core, staging]
  - name: raw

filters — narrow selection to matching tables

array of objects · optional

Each entry matches tables by name pattern, tag, or type. The type field is auto-inferred from the other fields present.

By name pattern:

Property	Type	Required	Description
`table_name`	string	yes	Pattern to match against table names
`table_name_operator`	string	yes	`STARTS_WITH` · `ENDS_WITH` · `CONTAINS` · `MATCH_PATTERN`

By tag:

Property	Type	Required	Description
`table_tags`	array of strings	yes	Tag values to match
`table_tags_operator`	string	yes	`HAS_ALL` · `HAS_ANY`

By type:

Property	Type	Required	Description
`table_type`	string	yes	`TABLE` · `VIEW` · `EXTERNAL`

filters:
  - table_name: fct_
    table_name_operator: STARTS_WITH
  - table_tags: [critical]
    table_tags_operator: HAS_ANY

exclusions — remove matching tables from selection

array of objects · optional

Same structure as filters. Tables matching any exclusion are removed even if they match a filter.

exclusions:
  - table_name: _deprecated
    table_name_operator: ENDS_WITH

alert_conditions — health signals to monitor

array of objects · optional · default: all metrics

Each entry has a single metric property. Table monitors do not accept operators or thresholds — all alerting uses anomaly detection.

`metric` value	What it detects
`last_updated_on`	Table not updated within expected cadence (freshness)
`total_row_count`	Unexpected row count changes (volume)
`total_row_count_last_changed_on`	Row count unchanged for unusual duration
`schema`	Any schema modification (superset of the three below)
`schema_fields_added`	New columns added
`schema_fields_removed`	Columns removed
`schema_fields_type_change`	Column data type changed

alert_conditions:
  - metric: last_updated_on
  - metric: total_row_count

domains — domain for this monitor

array of strings (exactly one entry) · required on accounts created after January 2025

Set default_domain in montecarlo.yml to avoid repeating it on every monitor.

domains:
  - my-domain

sensitivity — anomaly detection aggressiveness

enum · optional · default: account-level setting

The only tuning lever for table monitors — there are no explicit thresholds.

Value	Behavior
`low`	Fewer alerts, only large deviations
`medium`	Balanced
`high`	More alerts, smaller deviations flagged

sensitivity: medium

name — unique identifier within the namespace

string · required

Required for monitors created after Jan 29, 2024 (existing monitors keep working). Changing the name creates a new monitor and deletes the old one — incident history does not transfer.

name: analytics_freshness_check

warehouse — which warehouse to use

string · required if multiple warehouses

Warehouse UUID or name. Overrides default_resource from montecarlo.yml.

warehouse: my-snowflake

enable_row_count_collection — run explicit row count queries

boolean · optional · default: false

When enabled, Monte Carlo runs queries to count rows instead of relying on metadata. More accurate but uses warehouse compute.

Properties:

Property	Type	Required	Description
`enable_row_count_collection_limit`	integer	no	Max tables to query. Only applies when enabled.

enable_row_count_collection: true
enable_row_count_collection_limit: 500

tags — key-value pairs for organizing monitors

array of objects · optional

Property	Type	Required	Description
`name`	string	yes	Tag key
`value`	string	no	Tag value

tags:
  - name: team
    value: analytics
  - name: environment
    value: production

priority — incident priority level

enum · optional

Accepted values: P1 · P2 · P3 · P4 · P5

priority: P2

audiences — notification channels

array of strings · optional

Audience names linking this monitor to channels defined in Notifications as Code.

audiences:
  - data-engineering
  - platform-alerts

alert_grouping — control how breaches are grouped into alerts

object · optional · default: 5-hour schema-based grouping

Groups subsequent same-schema breaches into the currently open alert instead of the default 5-hour window. Freshness, volume, and time since last row count change alerts are always grouped together; schema change alerts group with other schema change alerts.

mode
Accepted values: group_into_open_alert

alert_grouping:
  mode: group_into_open_alert

data_quality_dimension — data quality category

enum · optional

Accepted values: ACCURACY · COMPLETENESS · CONSISTENCY · TIMELINESS · UNIQUENESS · VALIDITY

data_quality_dimension: TIMELINESS

notes — internal notes

string · optional

Visible in the Monte Carlo UI. Not included in notifications.

notes: Owned by the analytics team. Reviewed quarterly.

is_draft — create as draft without activating

boolean · optional · default: false

Creates the monitor in a paused state. Omitting this on a later update resets to false (active) due to PUT semantics — always include it if you want the monitor to stay in draft.

is_draft: true

uuid — update an existing monitor

string · optional

Include the UUID of an existing monitor to update it instead of creating a new one.

uuid: 0dae7702-0950-45c7-909c-7e183bddca19

Deprecated fields

Field	Use instead
`resource`	`warehouse`
`domain`	`domains`
`domain_uuids`	`domains`
`labels`	`audiences`

Examples

All tables in a database

montecarlo:
  table:
    - name: broad_analytics_monitoring
      description: Monitor all analytics tables
      asset_selection:
        databases:
          - name: analytics
      domains:
        - my-domain

Specific schemas with filtering

montecarlo:
  table:
    - name: core_fact_tables
      description: Monitor fact tables in core schema
      warehouse: my-snowflake
      asset_selection:
        databases:
          - name: analytics
            schemas: [core]
        filters:
          - table_name: fct_
            table_name_operator: STARTS_WITH
        exclusions:
          - table_name: _deprecated
            table_name_operator: ENDS_WITH
      alert_conditions:
        - metric: last_updated_on
        - metric: total_row_count
        - metric: schema
      sensitivity: medium
      audiences:
        - data-engineering
      priority: P2
      domains:
        - my-domain
      tags:
        - name: team
          value: analytics

Multiple databases with row count collection

montecarlo:
  table:
    - name: production_monitoring
      description: Cross-database production monitoring
      asset_selection:
        databases:
          - name: warehouse
            schemas: [public]
          - name: mart
            schemas: [finance, marketing]
      alert_conditions:
        - metric: last_updated_on
        - metric: total_row_count
        - metric: total_row_count_last_changed_on
      enable_row_count_collection: true
      enable_row_count_collection_limit: 500
      sensitivity: low
      domains:
        - my-domain

Troubleshooting

Scheduling and alerting

Adding a schedule field — table monitors run on Monte Carlo's collection cycle. There is no schedule field. Adding one causes a validation error.
Using operators or thresholds in alert_conditions — write { metric: last_updated_on }, not { metric: last_updated_on, operator: GT, threshold_value: 100 }. Table monitors use anomaly detection only.
Redundant schema metrics — schema is a superset of schema_fields_added, schema_fields_removed, and schema_fields_type_change. Using both is redundant.

Selection and updates

Attempting table-level selection — while databases[].tables exists in the schema, the backend rejects it. Use filters/exclusions for finer control.
Forgetting PUT semantics on updates — when updating a monitor by including uuid, every field you omit reverts to its default. Always specify the complete configuration, not just what changed.