Multi-App Observability Setup

Services that run multiple FastAPI apps (e.g. an API service on port 8000 and an upload service on port 8001) need each app independently instrumented. configure_app_observability does both steps — middleware and metrics endpoint — in a single call.

Quick Start

Python

from fastapi import FastAPI
from obskit import configure_observability
from obskit.middleware.instrument import configure_app_observability

obs = configure_observability(service_name="upload-service")

upload_app = FastAPI(title="upload-service")
configure_app_observability(upload_app, exclude_paths=["/v2/_healthy"])

This is equivalent to:

Python

from obskit import instrument_fastapi
from obskit.metrics.registry import generate_latest
from starlette.requests import Request
from starlette.responses import Response

instrument_fastapi(upload_app, exclude_paths=["/v2/_healthy"])

@upload_app.get("/metrics", include_in_schema=False)
async def metrics(request: Request) -> Response:
    return Response(generate_latest(), media_type="text/plain; version=0.0.4")

Parameters

Parameter	Default	Description
app	—	The FastAPI application.
exclude_paths	middleware defaults	Paths excluded from observability.
track_metrics	True	Enable RED metrics.
track_logging	True	Enable request/response logging.
track_tracing	True	Enable distributed tracing.
metrics_path	"/metrics"	URL path for the Prometheus scrape endpoint.

Custom Metrics Path

Python

configure_app_observability(upload_app, metrics_path="/internal/metrics")

API Reference

obskit.middleware.instrument.configure_app_observability

Python

configure_app_observability(
    app: Any,
    *,
    exclude_paths: list[str] | None = None,
    track_metrics: bool = True,
    track_logging: bool = True,
    track_tracing: bool = True,
    metrics_path: str = "/metrics",
    context_extractor: Any | None = None,
) -> None

Add :class:~obskit.middleware.fastapi.ObskitMiddleware and a Prometheus scrape endpoint to a FastAPI application in a single call.

Equivalent to calling :func:instrument_fastapi and manually registering a /metrics route — useful when running two or more FastAPI apps (e.g. an API service and an upload service) that each need full observability.

Parameters

app : FastAPI The FastAPI application to configure. exclude_paths : list[str], optional Paths excluded from middleware observability. Defaults to the :class:~obskit.middleware.core.MiddlewareCore defaults. track_metrics : bool Enable RED metrics collection. Default: True. track_logging : bool Enable structured request/response logging. Default: True. track_tracing : bool Enable distributed tracing. Default: True. metrics_path : str URL path for the Prometheus scrape endpoint. Default: "/metrics". context_extractor : callable, optional Forwarded to :class:~obskit.middleware.fastapi.ObskitMiddleware. Receives decoded request headers; returns extra log-context key/value pairs.

Example

::

Text Only

from obskit.middleware.instrument import configure_app_observability

upload_app = FastAPI(title="upload-service")
configure_app_observability(upload_app, exclude_paths=["/v2/_healthy"])
# ObskitMiddleware + GET /metrics are now registered on upload_app

Source code in src/obskit/middleware/instrument.py

def configure_app_observability(
    app: Any,
    *,
    exclude_paths: list[str] | None = None,
    track_metrics: bool = True,
    track_logging: bool = True,
    track_tracing: bool = True,
    metrics_path: str = "/metrics",
    context_extractor: Any | None = None,
) -> None:
    """Add :class:`~obskit.middleware.fastapi.ObskitMiddleware` and a Prometheus
    scrape endpoint to a FastAPI application in a single call.

    Equivalent to calling :func:`instrument_fastapi` and manually registering
    a ``/metrics`` route — useful when running two or more FastAPI apps (e.g.
    an API service and an upload service) that each need full observability.

    Parameters
    ----------
    app : FastAPI
        The FastAPI application to configure.
    exclude_paths : list[str], optional
        Paths excluded from middleware observability.  Defaults to the
        :class:`~obskit.middleware.core.MiddlewareCore` defaults.
    track_metrics : bool
        Enable RED metrics collection.  Default: ``True``.
    track_logging : bool
        Enable structured request/response logging.  Default: ``True``.
    track_tracing : bool
        Enable distributed tracing.  Default: ``True``.
    metrics_path : str
        URL path for the Prometheus scrape endpoint.  Default: ``"/metrics"``.
    context_extractor : callable, optional
        Forwarded to :class:`~obskit.middleware.fastapi.ObskitMiddleware`.
        Receives decoded request headers; returns extra log-context key/value pairs.

    Example
    -------
    ::

        from obskit.middleware.instrument import configure_app_observability

        upload_app = FastAPI(title="upload-service")
        configure_app_observability(upload_app, exclude_paths=["/v2/_healthy"])
        # ObskitMiddleware + GET /metrics are now registered on upload_app
    """
    instrument_fastapi(
        app,
        exclude_paths=exclude_paths,
        track_metrics=track_metrics,
        track_logging=track_logging,
        track_tracing=track_tracing,
        context_extractor=context_extractor,
    )

    from starlette.responses import Response  # noqa: PLC0415

    from obskit.metrics.registry import generate_latest  # noqa: PLC0415

    async def _metrics() -> Response:
        data = generate_latest()
        return Response(
            content=data,
            media_type="text/plain; version=0.0.4; charset=utf-8",
        )

    app.add_api_route(metrics_path, _metrics, include_in_schema=False)