prometheus-fastapi-instrumentator 1.2.0

Instrument your FastAPI with Prometheus metrics

Prometheus FastAPI Instrumentator

Instruments your FastAPI with Prometheus metrics. Install with:

pip install prometheus-fastapi-instrumentator

Fast Track

from prometheus_fastapi_instrumentator import Instrumentator
Instrumentator().instrument(app).expose(app)

With this single line FastAPI is instrumented and all Prometheus metrics used in the FastAPI app can be scraped via the added /metrics endpoint.

The exporter includes the single metric http_request_duration_seconds. Everything around it can be configured and deactivated. These options include:

• Status codes are grouped into 2xx, 3xx and so on.
• Requests without a matching template are grouped into the handler none.
• Renaming of labels and the metric.
• Regex patterns to ignore certain routes.
• Rounding of latencies.

See the Example with all parameters for all possible options or check out the documentation itself.

Example with all parameters

from prometheus_fastapi_instrumentator import PrometheusFastApiInstrumentator
PrometheusFastApiInstrumentator(
    should_group_status_codes=False,
    should_ignore_untemplated=True,
    should_group_untemplated=False,
    should_round_latency_decimals=True,
    excluded_handlers=["/metrics", "/admin"],
    buckets=[1, 2, 3, 4, 5],
    metric_name="my_custom_metric_name",
    label_names=("method_type", "path", "status_code",),
    round_latency_decimals=3,
).instrument(app).expose(app, "/prometheus_metrics")

instrument: Instruments the given FastAPI based on the configuration in the constructur of the exporter class.

expose: Completely separate from instrument and not necessary for instrumentation. Just a simple option to expose metrics by adding an endpoint to the given FastAPI. Supports multiprocess mode.

Prerequesites

• python = "^3.6" (tested with 3.6 and 3.8)
• fastapi = ">=0.38.1, <=1.0.0" (tested with 0.38.1 and 0.59.0)
• prometheus-client = "^0.8.0" (tested with 0.8.0)

Development

Developing and building this package on a local machine requires Python Poetry. I recommend to run Poetry in tandem with Pyenv. Once the repository is cloned, run poetry install and poetry shell. From here you may start the IDE of your choice.

For formatting, the black formatter is used. Run black . in the repository to reformat source files. It will respect the black configuration in the pyproject.toml.