Clustering service for managing homogeneous objects in OpenStack
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
senlin/doc/source/contributor/event_dispatcher.rst

4.6 KiB

Event Dispatchers

An event dispatcher<Dispatcher> is a processor that converts a given action in Senlin engine into certain format and then persists it into some storage or sends it to downstream processing software.

Since version 3.0.0, Senlin comes with some built-in dispatchers that can dump event records into database and/or send event notifications via the default message queue. The former is referred to as the database dispatcher which is enabled by default; the latter is referred to as the message dispatcher which has to be manually enabled by adding the following line to the senlin.conf file:

event_dispatchers = message

However, the distributors or the users can always add their own event dispatchers easily when needed.

Event dispatchers are managed as Senlin plugins. Once a new event dispatcher is implemented, a deployer can enable it by first adding a new item to the senlin.dispatchers entries in the entry_points section of the setup.cfg file, followed by a reinstall of the Senlin service, i.e. sudo pip install command.

The Base Class EventBackend

All event dispatchers are expected to subclass the base class EventBackend in the senlin.events.base module. The only requirement for a dispatcher subclass is to override the dump() method that implements the processing logic.

Providing New Dispatchers

Developing A New Event Dispatcher

The first step for adding a new dispatcher is to create a new file containing a subclass of EventBackend. In this new class, say JsonDispatcher, you will need to implement the dump() class method as exemplified below:

class JsonDispatcher(base.EventBackend):
    """Dispatcher for dumping events to a JSON file."""

    @classmethod
    def dump(cls, level, action, **kwargs):
        # Your logic goes here
        ...

The level parameter for the method is identical to that defined by the logging module of Python. It is an integer representing the criticality of an event. The action parameter is an instance of Senlin action class, which is defined in the senlin.engine.actions.base module. There is virtually no constraints on which properties you will pick and how you want to process them.

Finally, the **kwargs parameter may provide some useful fields for you to use:

  • timestamp: A datetime value that indicates when the event was generated.
  • phase: A string value indicating the phase an action is in. Most of the time this can be safely ignored.
  • reason: There are some rare cases where an event comes with a textual description. Most of the time, this is empty.
  • extra: There are even rarer cases where an event comes with some additional fields for attention. This can be safely ignored most of the time.

Registering the New Dispatcher

For Senlin service to be aware of and thus to make use of the new dispatcher, you will register it to the Senlin engine service. This is done by editing the setup.cfg file in the root directory of the code base, for example:

[entry_points]
senlin.dispatchers =
    database = senlin.events.database:DBEvent
    message = senlin.events.message:MessageEvent
    jsonfile = <path to the dispatcher module>:<dispatch class name>

Finally, save that file and do a reinstall of the Senlin service, followed by a restart of the senlin-engine process.

$ sudo pip install -e .

Dynamically Enabling/Disabling a Dispatcher

All dispatchers are loaded when the Senlin engine is started, however, they can be dynamically enabled or disabled by editing the senlin.conf file. The option event_dispatchers in the [DEFAULT] section is a multi-string value option for this purpose. To enable your dispatcher (i.e. jsonfile), you will need to add the following line to the senlin.conf file:

event_dispatchers = jsonfile