--- title: Custom Metrics in Database Monitoring description: >- Use the custom_queries option to collect metrics from your own database tables and view them in Database Monitoring. breadcrumbs: Docs > Database Monitoring > Custom Metrics in Database Monitoring --- > For the complete documentation index, see [llms.txt](https://docs.datadoghq.com/llms.txt). # Custom Metrics in Database Monitoring Use `custom_queries` to collect metrics from any table the Agent's database user can read. This extends the data available in Datadog beyond the query performance metrics that Database Monitoring collects natively, such as application state tables, business counters, or queue depths. ## Before you begin{% #before-you-begin %} The Datadog Agent must be installed and the database integration configured. The Agent's database user needs `SELECT` on any tables you query. ## Configuration{% #configuration %} Add `custom_queries` to your integration's `conf.yaml` file. Each entry in the list runs one SQL query and maps its output columns to metrics or tags. | Option | Required | Description | | --------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `metric_prefix` | No | All metrics emitted by this query begin with this prefix. | | `query` | Yes | The SQL to execute. All returned rows are evaluated. Use the pipe character (`|`) for multi-line queries. | | `columns` | Yes | A list of columns in the same order as your `SELECT`. Each column requires a `name` and a `type`. Set `type` to `gauge`, `count`, `rate`, or another [metric type](https://docs.datadoghq.com/metrics/types.md) to emit a metric, or `tag` to apply the column value as a tag on every metric from this query. | | `tags` | No | A list of static tags applied to every metric from this query. | | `collection_interval` | No | How often, in seconds, the Agent runs the query. Default: every 15 seconds. | If you set `metric_prefix`, metrics are named `.`. Otherwise, the integration's default prefix applies. For example, Postgres metrics are named `postgresql.`, and MySQL metrics are named ``. **Notes:** - The number of `columns` entries must equal the number of columns returned by the query. - The order of `columns` entries must match the order of columns returned by the query. - At least one entry in `columns` must be a metric type (not `tag`). ### Column types{% #column-types %} Each column in a custom query is assigned a type that controls how the metric is aggregated and displayed: | Type | Description | | ------------------ | --------------------------------------------------------- | | `gauge` | A value that can go up or down (for example, table size). | | `count` | A count of events since the last collection. | | `rate` | A per-second rate. | | `monotonic_count` | A counter that only increases. | | `monotonic_gauge` | A monotonically increasing gauge. | | `temporal_percent` | A percentage of time. | | `time_elapsed` | Duration in time units. | | `tag` | Groups or filters metrics; not plotted as its own graph. | `count` and `monotonic_count` columns are aggregated as `sum`. All other metric types are aggregated as `avg`. ## Examples{% #examples %} {% tab title="PostgreSQL" %} Add `custom_queries` to your `postgres.d/conf.yaml` file. If the query reads from a table the `datadog` user cannot already access, grant the permission first: ```sql GRANT SELECT ON TO datadog; ``` **Example:** The following `company` table contains employee records: ```text id | name | age | address | salary --------------------------------------- 1 | Paul | 32 | California | 20000 2 | Allen | 25 | Texas | 30000 3 | Teddy | 23 | Norway | 45000 ``` To collect `age` and `salary` as metrics with `name` and `address` as tags: ```yaml custom_queries: - metric_prefix: postgresql.employee query: SELECT age, salary, name, address FROM company columns: - name: employee_age type: gauge - name: employee_salary type: gauge - name: name type: tag - name: address type: tag tags: - source:hr_db ``` After you update the file, [restart the Agent](https://docs.datadoghq.com/agent/configuration/agent-commands.md#restart-the-agent). For the full configuration reference, see [Postgres Custom Metric Collection](https://docs.datadoghq.com/integrations/faq/postgres-custom-metric-collection-explained.md). {% /tab %} {% tab title="MySQL" %} Add `custom_queries` to your `mysql.d/conf.yaml` file. **Important:** All table references must include the database name (`database_name.table_name`). If you omit the database name, the Agent fails with the error: `No database selected`. **Example:** The following `company` table in the `testdb` database contains employee records: ```text id | name | age | address | salary --------------------------------------- 1 | Paul | 32 | California | 20000 2 | Allen | 25 | Texas | 30000 3 | Teddy | 23 | Norway | 45000 ``` To collect `age` and `salary` as metrics with `name` and `address` as tags: ```yaml custom_queries: - metric_prefix: mysql.employee query: SELECT age, salary, name, address FROM testdb.company columns: - name: employee_age type: gauge - name: employee_salary type: gauge - name: name type: tag - name: address type: tag tags: - source:hr_db ``` After you update the file, [restart the Agent](https://docs.datadoghq.com/agent/configuration/agent-commands.md#restart-the-agent). For the full configuration reference, see [MySQL Custom Queries](https://docs.datadoghq.com/integrations/guide/mysql-custom-queries.md). {% /tab %} {% tab title="SQL Server" %} SQL Server supports two approaches for collecting custom metrics: custom queries or performance counters. ### Custom queries{% #custom-queries %} Add `custom_queries` to your `sqlserver.d/conf.yaml` file to collect metrics from any table. **Example:** The following `company` table in `testdb` contains employee records: ```text id | name | age | address | salary --------------------------------------- 1 | Paul | 32 | California | 20000 2 | Allen | 25 | Texas | 30000 3 | Teddy | 23 | Norway | 45000 ``` To collect `age` and `salary` as metrics with `name` and `address` as tags: ```yaml custom_queries: - metric_prefix: sqlserver.employee query: SELECT age, salary, name, address FROM testdb.dbo.company columns: - name: employee_age type: gauge - name: employee_salary type: gauge - name: name type: tag - name: address type: tag tags: - source:hr_db ``` ### Performance counters{% #performance-counters %} Use `custom_metrics` to collect metrics from `sys.dm_os_performance_counters` and other system DMVs. ```yaml custom_metrics: - name: sqlserver.clr.execution counter_name: CLR Execution ``` | Option | Required | Description | | --------------- | -------- | ------------------------------------------------------------------------------------ | | `name` | Yes | The metric name in Datadog. | | `counter_name` | Yes | The counter name from `sys.dm_os_performance_counters`. | | `instance_name` | No | A specific counter instance. Use `ALL` to collect all instances (requires `tag_by`). | | `tag_by` | No | Tag name used to differentiate instances when `instance_name: ALL`. | After you update the file, [restart the Agent](https://docs.datadoghq.com/agent/configuration/agent-commands.md#restart-the-agent). For the full configuration reference, including performance counter details and the legacy stored procedure method, see [Collect SQL Server Custom Metrics](https://docs.datadoghq.com/integrations/guide/collect-sql-server-custom-metrics.md). {% /tab %} ## Validation{% #validation %} After the Agent runs, search for your metrics in the [Metrics Explorer](https://docs.datadoghq.com/metrics/explorer.md). To check for configuration errors, [run the Agent's status subcommand](https://docs.datadoghq.com/agent/configuration/agent-commands.md#agent-status-and-information) and look for your integration under the Checks section: ```text postgres -------- - instance #0 [ERROR]: 'Missing metric_prefix parameter in custom_queries' - Collected 0 metrics, 0 events & 0 service checks ``` ## Explore custom metrics{% #explore-custom-metrics %} The Custom Metrics section appears on the database instance detail page and displays timeseries graphs for any custom queries you have defined in your Datadog Agent configuration. If you have configured `custom_queries` in your Datadog Agent's database integration, this section automatically discovers those queries and visualizes each metric column as a timeseries graph. Graphs are scoped to the current database instance. This lets you monitor business-specific or environment-specific database metrics alongside the standard Database Monitoring metrics, all in one place. The Custom Metrics section is hidden if no custom queries are configured. ### Example Agent configuration{% #example-agent-configuration %} The following PostgreSQL example tracks table size and row counts per table: ```yaml init_config: instances: - dbm: true host: localhost port: 5432 username: datadog password: custom_queries: - metric_prefix: postgresql.custom query: | SELECT table_name, pg_total_relation_size(quote_ident(table_name)) AS total_bytes, n_live_tup AS live_rows, n_dead_tup AS dead_rows FROM information_schema.tables JOIN pg_stat_user_tables USING (table_name) WHERE table_schema = 'public' columns: - name: table_name type: tag - name: total_bytes type: gauge - name: live_rows type: gauge - name: dead_rows type: gauge collection_interval: 60 tags: - env:production - service:my-app ``` This configuration produces three metrics, each broken down by `table_name`: - `postgresql.custom.total_bytes` - `postgresql.custom.live_rows` - `postgresql.custom.dead_rows` All three appear as separate timeseries graphs in the Custom Metrics section of the instance detail page. ### View the source SQL{% #view-the-source-sql %} Each graph has a View SQL query button in the top-right corner. Clicking it shows the raw SQL statement that produces the metric, so you can understand and audit what is being measured. ### Collection interval{% #collection-interval %} The section subtitle shows how often the metrics are collected (for example, "collected every 15s"). If you have multiple custom queries with different intervals, the range is shown (for example, "collected every 15s-60s"). ## Further Reading{% #further-reading %} Additional helpful documentation, links, and articles: - [Database Monitoring](https://docs.datadoghq.com/database_monitoring.md) - [Metric Types](https://docs.datadoghq.com/metrics/types.md)