temporalio.worker.Plugin

class documentation

class Plugin(abc.ABC): (source)

Known subclasses: temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Base class for worker plugins that can intercept and modify worker behavior.

Plugins allow customization of worker creation and execution processes through a chain of responsibility pattern. Each plugin can modify the worker configuration or intercept worker execution.

WARNING: This is an experimental feature and may change in the future.

Method	`configure_replayer`	Hook called when creating a replayer to allow modification of configuration.
Method	`configure_worker`	Hook called when creating a worker to allow modification of configuration.
Method	`init_worker_plugin`	Initialize this plugin in the plugin chain.
Method	`name`	Get the qualified name of this plugin. Can be overridden if desired to provide a more appropriate name.
Method	`run_replayer`	Hook called when running a replayer to allow interception of execution.
Async Method	`run_worker`	Hook called when running a worker to allow interception of execution.

@abc.abstractmethod
def configure_replayer(self, config: ReplayerConfig) -> ReplayerConfig: (source) ¶

overridden in temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Hook called when creating a replayer to allow modification of configuration.

This should be used to configure anything in ReplayerConfig needed to make execution match the worker and client config. This could include interceptors, DataConverter, workflows, and more.

Parameters
config:`ReplayerConfig`	The replayer configuration dictionary to potentially modify.
Returns
`ReplayerConfig`	The modified replayer configuration.

@abc.abstractmethod
def configure_worker(self, config: WorkerConfig) -> WorkerConfig: (source) ¶

overridden in temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Hook called when creating a worker to allow modification of configuration.

This method is called during worker creation and allows plugins to modify the worker configuration before the worker is fully initialized. Plugins can modify task queue names, adjust concurrency settings, add interceptors, or change other worker settings.

Parameters
config:`WorkerConfig`	The worker configuration dictionary to potentially modify.
Returns
`WorkerConfig`	The modified worker configuration.

@abc.abstractmethod
def init_worker_plugin(self, next: Plugin): (source) ¶

overridden in temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Initialize this plugin in the plugin chain.

This method sets up the chain of responsibility pattern by providing a reference to the next plugin in the chain. It is called during worker creation to build the plugin chain. Implementations should store this reference and call the corresponding method of the next plugin on method calls.

Parameters
next:`Plugin`	The next plugin in the chain to delegate to.

def name(self) -> str: (source) ¶

Get the qualified name of this plugin. Can be overridden if desired to provide a more appropriate name.

Returns
`str`	The fully qualified name of the plugin class (module.classname).

@abc.abstractmethod
def run_replayer(self, replayer: Replayer, histories: AsyncIterator[WorkflowHistory]) -> AbstractAsyncContextManager[AsyncIterator[WorkflowReplayResult]]: (source) ¶

overridden in temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Hook called when running a replayer to allow interception of execution.

@abc.abstractmethod
async def run_worker(self, worker: Worker): (source) ¶

overridden in temporalio.contrib.openai_agents.OpenAIAgentsPlugin, temporalio.worker._plugin._RootPlugin

Hook called when running a worker to allow interception of execution.

This method is called when the worker is started and allows plugins to intercept or wrap the worker execution. Plugins can add monitoring, custom lifecycle management, or other execution-time behavior.

Parameters
worker:`Worker`	The worker instance to run.