Rulebooks

Rulebooks contain a list of rulesets. Each ruleset within a rulebook

should have a unique name since they can post events to each other at runtime

based on the name. If a rulebook has multiple rulesets

shutting down one ruleset will shutdown all the other running rulesets.

Rulesets

A ruleset has the following properties:

Name	Description	Required
name	The name to identify the ruleset. Each ruleset must have an unique name across the rulebook.	Yes
sources	The list of one or more sources that will generate events for ansible-rulebook. See Event Source Plugins	Yes
rules	The list of one or more rule. See Rules	Yes
hosts	Similar to hosts in an Ansible playbook	Yes
gather_facts	Collect artifacts from hosts at startup to be used in rules (default: false)	No
default_events_ttl	time to keep the partially matched events around (default: 2 hours)	No
execution_strategy	Action execution, sequential or parallel (default: sequential). For sequential strategy we wait for the each action to finish before firing of the next action.	No

A ruleset should have a unique name within the rulebook, each ruleset runs

as a separate session in the Rules engine. The events and facts are kept separate

for each ruleset. At runtime, using action a ruleset can post events or facts

to itself or other rulesets in the rulebook.

The default_events_ttl takes time in the following format

default_events_ttl : nnn seconds|minutes|hours|days

e.g. default_events_ttl : 3 hours

If the rule set doesn’t define this attribute the default events ttl that is

enforced by the rule engine is 2 hours

When we start a rulebook we can optionally collect artifacts from the different hosts

if gather_facts is set to true. This host data is then uploaded to the Rules

engine as fact to be evaluated at runtime in the different rules based on the

incoming events. Each host data is stored separately in the Rules engine. To access the

host name use the fact.meta.hosts attribute. e.g.

- name: Example
  hosts: all
  gather_facts: true
  sources:
    - name: range
      ansible.eda.range:
        limit: 5
  rules:
    - name: r1
      condition: event.i == 1
      action:
        debug:

    - name: "Host specific rule"
      condition:
        all:
          - fact.ansible_os_family == "linux"
          - fact.meta.hosts == "my-host"
          - event.i == 4
      action:
        debug:

A ruleset must contain one or more source plugins, the configuration parameters

can be specified after the source plugin type. The source plugin

can also be configured with event filters which allow you to transform the

data before passing it to the rules engine. The filters can also be used to

limit the data that gets passed to the rules engine. The source plugin is

started by the ansible-rulebook and runs in the background putting events

into the queue to be passed onto the rules engine.

When the source plugin ends we automatically generate a shutdown event and the ruleset

terminates which terminates ansible-rulebook.

A ruleset must contain one or more rules. The rules are evaluated by the Rules engine.

The Rules engine will evaluate all the required conditions for a rule based on the

incoming events. If the conditions in a rule match, we trigger the actions. The actions

can run playbooks, modules, raise another event or fact to the same ruleset or a different

ruleset. A ruleset stops execution when it receives the shutdown event from either the

Source plugin or a shutdown action is invoked by one of the matching rules.

Including multiple sources

In a rulebook you can configure one or more sources, each emitting events in different format.

Example

sources:
  - ansible.eda.range:
      limit: 6
  - ansible.eda.webhook:
      port: 5000

The condition can match events from either source

rules:
  - name:
    condition: event.i == 2
    action:
      debug:

  - name:
    condition: event.payload.status == "OK"
    action:
      debug:

To avoid name conflicts the source data structure can use nested keys.

Notes:

If any source terminates, it shuts down the whole engine. All events from other sources may be lost.

Distributing rulebooks

The recommended method for distributing rulebooks is through a collection. In this case the rulebook file should be placed under extensions/eda/rulebooks folder and referred to by FQCN in the command line argument. Eda-server project will honor this path for the projects even if the repository is not real collection.