temporalio.contrib.gcp.cloud_run.opentelemetry.OpenTelemetryPlugin

class documentation

class OpenTelemetryPlugin(TemporalOpenTelemetryPlugin): (source)

Constructor: OpenTelemetryPlugin(endpoint, service_name, metric_periodicity, flush_timeout, ...)

OpenTelemetry plugin for Temporal workers running on Google Cloud Run.

The default configuration sends Temporal Core metrics and Python traces over OTLP gRPC to a collector on localhost:4317. The collector is responsible for Google Cloud resource detection, authentication, and export.

The plugin creates and installs a replay-safe global tracer provider unless tracer_provider is supplied. It also creates a Temporal runtime for Core metrics unless runtime is supplied. Call shutdown after every worker using the plugin has stopped.

Warning

This class is experimental and may change in future versions. Use with caution in production environments.

Method	`__init__`	Initialize the Google Cloud OpenTelemetry plugin.
Async Method	`connect_service_client`	Install the metrics runtime before connecting the service client.
Method	`force_flush`	Export buffered Python traces without shutting down the provider.
Async Method	`run_worker`	Run a worker and optionally force-flush traces after it stops.
Method	`shutdown`	Flush traces and shut down a provider created by the plugin.
Property	`endpoint`	Resolved OTLP collector endpoint.
Property	`runtime`	Temporal runtime used for Core metrics.
Property	`service_name`	Resolved OpenTelemetry service name.
Property	`tracer_provider`	Replay-safe tracer provider used by the plugin.
Instance Variable	`_endpoint`	Undocumented
Instance Variable	`_flush_on_worker_stop`	Undocumented
Instance Variable	`_flush_timeout`	Undocumented
Instance Variable	`_owns_tracer_provider`	Undocumented
Instance Variable	`_runtime`	Undocumented
Instance Variable	`_service_name`	Undocumented
Instance Variable	`_shutdown`	Undocumented
Instance Variable	`_shutdown_lock`	Undocumented
Instance Variable	`_shutdown_succeeded`	Undocumented
Instance Variable	`_tracer_provider`	Undocumented

Inherited from SimplePlugin (via OpenTelemetryPlugin):

Method	`configure_client`	See base class.
Method	`configure_replayer`	See base class.
Method	`configure_worker`	See base class.
Method	`name`	See base class.
Async Method	`run_replayer`	See base class.
Instance Variable	`activities`	Undocumented
Instance Variable	`data_converter`	Undocumented
Instance Variable	`interceptors`	Undocumented
Instance Variable	`nexus_service_handlers`	Undocumented
Instance Variable	`run_context`	Undocumented
Instance Variable	`workflow_failure_exception_types`	Undocumented
Instance Variable	`workflow_runner`	Undocumented
Instance Variable	`workflows`	Undocumented
Instance Variable	`_name`	Undocumented

def __init__(self, *, endpoint: str | None = None, service_name: str | None = None, metric_periodicity: timedelta | None = None, flush_timeout: timedelta = DEFAULT_FLUSH_TIMEOUT, flush_on_worker_stop: bool = False, tracer_provider: TracerProvider | None = None, runtime: Runtime | None = None, add_temporal_spans: bool = False): (source) ¶

overrides temporalio.contrib.opentelemetry.OpenTelemetryPlugin.__init__

Initialize the Google Cloud OpenTelemetry plugin.

Parameters
endpoint:`str \| None`	OTLP gRPC endpoint. Falls back to `OTEL_EXPORTER_OTLP_ENDPOINT`, then `http://localhost:4317`.
service_name:`str \| None`	OpenTelemetry service name. Falls back to `OTEL_SERVICE_NAME`, `CLOUD_RUN_WORKER_POOL`, `K_SERVICE`, then `temporal-worker`.
metric_periodicity:`timedelta \| None`	How often Temporal Core metrics are exported. Defaults to 60 seconds. Cannot be used with `runtime`.
flush_timeout:`timedelta`	Default tracing force-flush timeout.
flush_on_worker_stop:`bool`	Whether to force-flush traces after each worker stops. Disabled by default because one plugin can be used by multiple workers.
tracer_provider:`TracerProvider \| None`	Application-owned replay-safe tracer provider. It must have been created with `temporalio.contrib.opentelemetry.create_tracer_provider`. When supplied, the plugin does not create an exporter or shut down the provider.
runtime:`Runtime \| None`	Application-owned Temporal runtime. When supplied, the plugin does not create or modify Core metrics configuration. Use `build_metrics_telemetry_config` to build a composable telemetry configuration for a custom runtime.
add_temporal_spans:`bool`	Whether the underlying Temporal OpenTelemetry plugin should add Temporal-specific operation spans.

async def connect_service_client(self, config: ConnectConfig, next: Callable[[ConnectConfig], Awaitable[ServiceClient]]) -> ServiceClient: (source) ¶

overrides temporalio.plugin.SimplePlugin.connect_service_client

Install the metrics runtime before connecting the service client.

def force_flush(self, timeout: timedelta | None = None) -> bool: (source) ¶

Export buffered Python traces without shutting down the provider.

Temporal Core metrics are exported periodically and the Python runtime currently has no explicit metrics-flush API.

Parameters
timeout:`timedelta \| None`	Maximum time to wait. Defaults to `flush_timeout` from the constructor.

Returns
`bool`	`True` when the tracer provider reports a successful flush.

async def run_worker(self, worker: Worker, next: Callable[[Worker], Awaitable[None]]): (source) ¶

overrides temporalio.plugin.SimplePlugin.run_worker

Run a worker and optionally force-flush traces after it stops.

def shutdown(self, timeout: timedelta | None = None) -> bool: (source) ¶

Flush traces and shut down a provider created by the plugin.

Application-owned tracer providers are force-flushed but are not shut down. This method is idempotent. Stop every worker using the plugin before calling it.