Monitoring & Activity Tracking

Metatate provides built-in observability through an append-only activity log, governance snapshots, a decision log, and standard SPCS container monitoring.

Activity Log

Table: app_data.activity_log

An append-only audit trail that records every significant action in the application.

Schema

Column	Type	Description
id	INTEGER	Auto-increment primary key
action	VARCHAR	Action identifier (e.g., policy.created, compile.applied, classification.run)
entity_type	VARCHAR	Type of entity affected (policy, table, taxonomy, instruction)
entity_id	VARCHAR	Unique identifier of the affected entity
entity_name	VARCHAR	Human-readable name of the entity
actor	VARCHAR	Snowflake user who performed the action
actor_role	VARCHAR	Snowflake role active at the time of the action
details	VARIANT	JSON payload with action-specific metadata
created_at	TIMESTAMP_NTZ	Timestamp of the event

API Access

GET /v1/activities

Query parameters:

Parameter	Type	Default	Description
limit	integer	50	Maximum number of records to return
offset	integer	0	Number of records to skip for pagination
action	string	--	Filter by action identifier (exact match)
entityType	string	--	Filter by entity type (exact match)

Governance Snapshots

Table: app_data.governance_snapshots

Tracks each materialization run that compiles policies into the governance tables consumed by MCP tools. Each row represents a point-in-time snapshot of the governance state, enabling diff-based auditing of what changed between materializations.

Decision Log

Table: app_data.governance_decision_log

Records every authorization decision made by the authorize_use and validate_query_context MCP tools. Each entry includes the request parameters, the decision (allow/deny), the reasoning chain, and the policies evaluated. Decisions are referenced by a unique decision_id that can be passed to the explain_why MCP tool for detailed review.

SPCS Container Logs

Container logs for all three services (router, frontend, backend) are accessible through Snowflake's built-in SPCS log infrastructure:

-- View recent backend logs
SELECT *
FROM TABLE(
  SYSTEM$GET_SERVICE_LOGS(
    'core.metatate_service', 0, 'backend', 100
  )
);

-- View router (nginx) logs
SELECT *
FROM TABLE(
  SYSTEM$GET_SERVICE_LOGS(
    'core.metatate_service', 0, 'router', 50
  )
);

Logs are retained according to the compute pool's configured log retention period.

Health Check

GET /health

Returns the current service status, version, and connectivity state. Used by SPCS as the container liveness probe and by the frontend to display connection status.

Example response:

{
  "status": "healthy",
  "version": "1.0.0",
  "snowflake_connected": true,
  "tenant_configured": true
}