+Create a monitor using the specified options. #### Monitor Types The type of monitor chosen from: - anomaly: `query alert` - APM: `query alert` or `trace-analytics alert` - composite: `composite` - custom: `service check` - event: `event alert` - forecast: `query alert` - host: `service check` - integration: `query alert` or `service check` - live process: `process alert` - logs: `log alert` - metric: `metric alert` - network: `service check` - outlier: `query alert` - process: `service check` - rum: `rum alert` - SLO: `slo alert` - watchdog: `event alert` - event-v2: `event-v2 alert` #### Query Types **Metric Alert Query** Example: `time_aggr(time_window):space_aggr:metric{tags} [by {key}] operator #` - `time_aggr`: avg, sum, max, min, change, or pct_change - `time_window`: `last_#m` (with `#` between 1 and 2880 depending on the monitor type) or `last_#h`(with `#` between 1 and 48 depending on the monitor type), or `last_1d` - `space_aggr`: avg, sum, min, or max - `tags`: one or more tags (comma-separated), or * - `key`: a 'key' in key:value tag syntax; defines a separate alert for each tag in the group (multi-alert) - `operator`: <, <=, >, >=, ==, or != - `#`: an integer or decimal number used to set the threshold If you are using the `_change_` or `_pct_change_` time aggregator, instead use `change_aggr(time_aggr(time_window), timeshift):space_aggr:metric{tags} [by {key}] operator #` with: - `change_aggr` change, pct_change - `time_aggr` avg, sum, max, min [Learn more](https://docs.datadoghq.com/monitors/monitor_types/#define-the-conditions) - `time_window` last\\_#m (between 1 and 2880 depending on the monitor type), last\\_#h (between 1 and 48 depending on the monitor type), or last_#d (1 or 2) - `timeshift` #m_ago (5, 10, 15, or 30), #h_ago (1, 2, or 4), or 1d_ago Use this to create an outlier monitor using the following query: `avg(last_30m):outliers(avg:system.cpu.user{role:es-events-data} by {host}, 'dbscan', 7) > 0` **Service Check Query** Example: `\"check\".over(tags).last(count).by(group).count_by_status()` - **`check`** name of the check, e.g. `datadog.agent.up` - **`tags`** one or more quoted tags (comma-separated), or \"*\". e.g.: `.over(\"env:prod\", \"role:db\")`; **`over`** cannot be blank. - **`count`** must be at greater than or equal to your max threshold (defined in the `options`). It is limited to 100. For example, if you've specified to notify on 1 critical, 3 ok, and 2 warn statuses, `count` should be at least 3. - **`group`** must be specified for check monitors. Per-check grouping is already explicitly known for some service checks. For example, Postgres integration monitors are tagged by `db`, `host`, and `port`, and Network monitors by `host`, `instance`, and `url`. See [Service Checks](https://docs.datadoghq.com/api/latest/service-checks/) documentation for more information. **Event Alert Query** Example: `events('sources:nagios status:error,warning priority:normal tags: \"string query\"').rollup(\"count\").last(\"1h\")\"` - **`event`**, the event query string: - **`string_query`** free text query to match against event title and text. - **`sources`** event sources (comma-separated). - **`status`** event statuses (comma-separated). Valid options: error, warn, and info. - **`priority`** event priorities (comma-separated). Valid options: low, normal, all. - **`host`** event reporting host (comma-separated). - **`tags`** event tags (comma-separated). - **`excluded_tags`** excluded event tags (comma-separated). - **`rollup`** the stats roll-up method. `count` is the only supported method now. - **`last`** the timeframe to roll up the counts. Examples: 45m, 4h. Supported timeframes: m, h and d. This value should not exceed 48 hours. **NOTE** Only available on US1 and EU. **Event V2 Alert Query** Example: `events(query).rollup(rollup_method[, measure]).last(time_window) operator #` - **`query`** The search query - following the [Log search syntax](https://docs.datadoghq.com/logs/search_syntax/). - **`rollup_method`** The stats roll-up method - supports `count`, `avg` and `cardinality`. - **`measure`** For `avg` and cardinality `rollup_method` - specify the measure or the facet name you want to use. - **`time_window`** #m (5, 10, 15, or 30), #h (1, 2, or 4, 24). - **`operator`** `<`, `<=`, `>`, `>=`, `==`, or `!=`. - **`#`** an integer or decimal number used to set the threshold. **NOTE** Only available on US1-FED, US3, and in closed beta on EU and US1. **Process Alert Query** Example: `processes(search).over(tags).rollup('count').last(timeframe) operator #` - **`search`** free text search string for querying processes. Matching processes match results on the [Live Processes](https://docs.datadoghq.com/infrastructure/process/?tab=linuxwindows) page. - **`tags`** one or more tags (comma-separated) - **`timeframe`** the timeframe to roll up the counts. Examples: 10m, 4h. Supported timeframes: s, m, h and d - **`operator`** <, <=, >, >=, ==, or != - **`#`** an integer or decimal number used to set the threshold **Logs Alert Query** Example: `logs(query).index(index_name).rollup(rollup_method[, measure]).last(time_window) operator #` - **`query`** The search query - following the [Log search syntax](https://docs.datadoghq.com/logs/search_syntax/). - **`index_name`** For multi-index organizations, the log index in which the request is performed. - **`rollup_method`** The stats roll-up method - supports `count`, `avg` and `cardinality`. - **`measure`** For `avg` and cardinality `rollup_method` - specify the measure or the facet name you want to use. - **`time_window`** #m (between 1 and 2880), #h (between 1 and 48) - **`operator`** `<`, `<=`, `>`, `>=`, `==`, or `!=`. - **`#`** an integer or decimal number used to set the threshold. **Composite Query** Example: `12345 && 67890`, where `12345` and `67890` are the IDs of non-composite monitors * **`name`** [*required*, *default* = **dynamic, based on query**]: The name of the alert. * **`message`** [*required*, *default* = **dynamic, based on query**]: A message to include with notifications for this monitor. Email notifications can be sent to specific users by using the same '@username' notation as events. * **`tags`** [*optional*, *default* = **empty list**]: A list of tags to associate with your monitor. When getting all monitor details via the API, use the `monitor_tags` argument to filter results by these tags. It is only available via the API and isn't visible or editable in the Datadog UI. **SLO Alert Query** Example: `error_budget(\"slo_id\").over(\"time_window\") operator #` - **`slo_id`**: The alphanumeric SLO ID of the SLO you are configuring the alert for. - **`time_window`**: The time window of the SLO target you wish to alert on. Valid options: `7d`, `30d`, `90d`. - **`operator`**: `>=` or `>`
0 commit comments