Rules

This content applies to:

In general, a rule consists of three parts:

For example, when monitoring a set of servers, a rule might:

Each project type supports a specific set of rule types. Each rule type provides its own way of defining the conditions to detect, but an expression formed by a series of clauses is a common pattern. For example, in an Elasticsearch query rule, you specify an index, a query, and a threshold, which uses a metric aggregation operation (count, average, max, min, or sum):

All rules must have a check interval, which defines how often to evaluate the rule conditions. Checks are queued; they run as close to the defined value as capacity allows.

You can add one or more actions to your rule to generate notifications when its conditions are met. Recovery actions likewise run when rule conditions are no longer met.

When defining actions in a rule, you specify:

Each action uses a connector, which provides connection information for a Kibana service or third party integration, depending on where you want to send the notifications. The specific list of connectors that you can use in your rule vary by project type. Refer to Connectors.

After you select a connector, set the action frequency. If you want to reduce the number of notifications you receive without affecting their timeliness, some rule types support alert summaries. For example, if you create an Elasticsearch query rule, you can set the action frequency such that you receive summaries of the new, ongoing, and recovered alerts on a custom interval:

Alternatively, you can set the action frequency such that the action runs for each alert. If the rule type does not support alert summaries, this is your only available option. You must choose when the action runs (for example, at each check interval, only when the alert status changes, or at a custom action interval). You must also choose an action group, which affects whether the action runs. Each rule type has a specific set of valid action groups. For example, you can set Run when to Query matched or Recovered for the Elasticsearch query rule:

Each connector supports a specific set of actions for each action group and enables different action properties. For example, you can have actions that create an Opsgenie alert when rule conditions are met and recovery actions that close the Opsgenie alert.

Some types of rules enable you to further refine the conditions under which actions run. For example, you can specify that actions run only when an alert occurs within a specific time frame or when it matches a KQL query.

You can pass rule values to an action at the time a condition is detected. To view the list of variables available for your rule, click the "add rule variable" button:

For more information about common action variables, refer to Rule actions variables

When checking for a condition, a rule might identify multiple occurrences of the condition. Kibana tracks each of these alerts separately. Depending on the action frequency, an action occurs per alert or at the specified alert summary interval.

Using the server monitoring example, each server with average CPU > 0.9 is tracked as an alert. This means a separate email is sent for each server that exceeds the threshold whenever the alert status changes.

A rule consists of conditions, actions, and a schedule. When conditions are met, alerts are created that render actions and invoke them. To make action setup and update easier, actions use connectors that centralize the information used to connect with Kibana services and third-party integrations. The following example ties these concepts together: