Cloudera Docs

Triggers

A trigger is a statement that specifies an action to be taken when one or more specified conditions are met for a service, role, role configuration group, or host. The conditions are expressed as a tsquery statement, and the action to be taken is to change the health for the service, role, role configuration group, or host to either Concerning (yellow) or Bad (red).

Triggers can be created for services, roles, role configuration groups, or hosts. Create a trigger by doing one of the following:
  • Directly editing the configuration for the service, role (or role configuration group), or host configuration.
  • Clicking Create Trigger on the drop-down menu for most charts. Note that the Create Trigger command is not available on the drop-down menu for charts where no context (role, service, and so on) is defined, such as on the Home > Status tab.
  • Use the Create Trigger expression builder.

The Structure of Triggers

A trigger is defined by a JSON formatted string that includes four parts:
  • Name
  • Expression
  • Stream threshold
  • Whether or not the trigger should be enabled
Each of the four parts of a trigger is described in the following sections.

Name (required)

A trigger's name must be unique in the context for which the trigger is defined. That is, there cannot be two triggers for the same service or role with the same name. Different services or different roles can have triggers with the same name.

Expression (required)

A trigger expression takes the form:

IF (CONDITIONS) DO HEALTH_ACTION

When the conditions of the trigger are met, the trigger is considered to be firing. A condition is any valid tsquery statement. In most cases conditions employ stream filters to filter out streams below or above a certain threshold. For example, the following tsquery can be used to retrieve the streams for DataNodes with more than 500 open file descriptors:
SELECT fd_open WHERE roleType=DataNode AND last(fd_open) > 500
The stream filter used here, last(fd_open) > 50, is composed of four parts:
  • A scalar producing function "last" that takes a stream and returns its last data point
  • A metric to operate on
  • A comparator
  • A scalar value
Other scalar producing functions are available, like min or max, and they can be combined to create arbitrarily complex expressions:
last(moving_avg(fd_open)) >= 500

See the tsquery documentation for more details.

Conditions can be combined using the logical operators AND and OR. For example, here is a trigger expression with two conditions:
IF ((SELECT fd_open WHERE roleType=DataNode AND last(fd_open) > 500) OR (SELECT fd_open WHERE roleType=NameNode AND last(fd_open) > 500)) DO health:bad

A condition is met if it returns more than the number of streams specified by the streamThreshold (see below). A trigger fires if the logical evaluation of all of its conditions results in a met condition. When a trigger fires, two actions can be taken: health:concerning or health:bad. These actions change the health of the entity on which the trigger is defined.

Stream Threshold (optional)

The stream threshold determines the number of streams that need to be returned by the tsquery before the condition is met. The default is 0; that is, if the tsquery returns any results the condition will be met. For example if the stream threshold is set to 10 and the condition is SELECT fd_open WHERE roleType=DataNode AND last(fd_open) > 500 the condition will be considered met only if there are at least 10 DataNodes that have more than 500 file descriptor opened, so at least 10 streams were returned by the tsquery.

Enabled (optional)

Whether the trigger is enabled. The default is true, (enabled).

Trigger Example

The following is a JSON formatted trigger that fires if there are more than 10 DataNodes with more than 500 file descriptors opened: 
[{"triggerName": "sample-trigger", "triggerExpression": "IF (SELECT fd_open WHERE roleType = DataNode and last(fd_open) > 500) DO health:bad", "streamThreshold": 10, "enabled": "true"}]