ℹ️Composable dashboards and the metrics API are in private preview: available to invited accounts, and features may change. See Integration and feature lifecycles for what this means, and reach out to your account representative to request access.
🤖
Dimensions are what you group and filter metrics by. They are listed by what they describe, such as a table, a monitor or an alert, and their names start with it: monitor__monitor_type is the type of a monitor. Grouping by monitor itself gives one row per monitor. Which dimensions a metric groups by is on its row under Metrics.
- Tables: Filter
asset__asset_category is table for tables and views, without AI agents; asset__asset_type tells tables from views.
- Names: Two monitors or data products can share a name; group by
monitor, data_product or domain itself and show its name next to it. Schemas can share a name across databases: group by asset__asset_project_name and asset__asset_dataset_name together. A table's full name (asset__asset_full_table_id) can be in several warehouses: group by asset and show the name next to it.
- Totals across members: A table, monitor or alert can belong to several domains, data products, asset tags, monitor domains, monitor tags or audiences, or to none, so the members' numbers do not add up to the account's. Show the account's number, from the same metric without the grouping, next to them; a member's share of it (%) says the shares overlap. Groupings by something each item has once (a monitor's type, an alert's severity, a warehouse) do add up.
- Rates across members: An average of the members' rates is not the account's rate; give the account's rate.
- Combining groupings: A per-grouping metric cannot be split by another grouping or by table. Take two steps: list the member's tables (
asset__asset_mcon) or monitors (monitor__monitor_uuid) first, then filter the other query on them.
- Warehouse: A table, a monitor, an alert and a credit each have their own warehouse;
warehouse__warehouse_name takes the one of the metric's own item. When a request refuses it as ambiguous, put the item in front of it: asset__ for the table's warehouse, monitor__ for the monitor's; the answer can name the column that way too. A warehouse that only monitors use has no tables, so it is in monitor and credit numbers, not in table numbers.
- Tags: Asset tags (
asset_tag__asset_tag_name) are on tables, monitor tags (monitor_tag__monitor_tag_name) on monitors and the alerts they raise.
- Deleted items: Deleted monitors and data products stay in the data; filter
monitor__is_monitor_deleted or data_product__is_data_product_deleted is false to leave them out.
- Data products: A data product's assets are the tables and reports added to it plus the tables upstream of them in lineage;
data_product_asset__is_data_product_root_asset keeps the added ones.
| Dimensions | What it is |
|---|
| Asset | A table, view or other asset Monte Carlo discovered in a connected source, or an AI agent; asset is its unique id (MCON). |
| Warehouse | A warehouse or other data source connected to the account. |
| Asset tag | A key:value label on tables, set in Monte Carlo or imported from the source. Tags sent through the ingestion API and use-case tags are not included. |
| Dimensions | What it is |
|---|
| Monitor | A check Monte Carlo runs on data or AI agents; it raises alerts when it finds a problem. Generated rules and bulk children have their own key and point to their parent (monitor__parent_monitor_uuid). |
| Monitor tag | A key:value label on monitors, often used for team ownership. |
| Monitor domain | The domain a monitor is assigned to: an attribute of the monitor, set on it in Monte Carlo, separate from the domains of the tables it checks. Alerts take the domain of the monitor that raised them; alerts without a monitor take the domains of their tables. |
| Monitor run | A run of a monitor: one point on its chart. Group by it to get one row per run. |
| Monitored metric | What a monitor measures: its metric, field and condition. |
| Segment | One segment of what a monitor measures. |
| Dimensions | What it is |
|---|
| Alert | An issue a monitor raised, grouping one or more events. Group by it to list alerts. |
| Alert sub type | What raised the alert, at the finest level (volume_anomaly, data_added, ...); one alert type covers one or more of them. In the product: the sub_types values of a link to the Alerts page; the Alert type filter shows incident_sub_type__incident_alert_type. |
| Event | One detection on a table (anomaly, schema change, rule breach); one or more events make an alert. |
| Alert activity | An action on an alert: a status, owner, severity or name change, a comment, a split, a merge or detector feedback (mark as normal). |
| Audience | A named group of notification recipients that alerts are routed to. |
| Dimensions | What it is |
|---|
| Account | Your Monte Carlo account (tenant). Every answer is for your account only: the API adds the account filter to every query. |
| Domain | A named set of tables used to scope access and reporting. |
| Data product | The tables and reports a user added to it and the tables upstream of them. |
| User | A Monte Carlo user of your account. |
| Dimensions | What it is |
|---|
| Consumption | Credits consumed: one row per monitor or agent, charge type and day. |
