obskit

Production-ready observability toolkit for Python microservices.

obskit is a single package with optional extras that brings structured logging, Prometheus metrics, OpenTelemetry tracing, and health checks to your Python services — with a consistent API and zero boilerplate.

Features at a glance

Feature	Extra	Install
Structured logging + trace correlation	(core, always included)	pip install obskit
Sensitive field redaction (structlog processor)	(core, always included)	pip install obskit
RED metrics + exemplars	prometheus	pip install "obskit[prometheus]"
Distributed tracing (OTel)	otlp	pip install "obskit[otlp]"
FastAPI middleware	fastapi	pip install "obskit[fastapi]"
Flask middleware	flask	pip install "obskit[flask]"
Django middleware	django	pip install "obskit[django]"
Health check framework	health	pip install "obskit[health]"
Health check framework + HTTP checks	health-http	pip install "obskit[health-http]"
SLO tracking	slo	pip install "obskit[slo]"
SLO tracking + Prometheus burn-rate metrics	slo-prometheus	pip install "obskit[slo-prometheus]"
SQLAlchemy query tracking	sqlalchemy	pip install "obskit[sqlalchemy]"
psycopg2 instrumentation	psycopg2	pip install "obskit[psycopg2]"
psycopg3 instrumentation	psycopg3	pip install "obskit[psycopg3]"
All DB drivers	db	pip install "obskit[db]"
Kafka instrumentation	kafka	pip install "obskit[kafka]"
RabbitMQ instrumentation	rabbitmq	pip install "obskit[rabbitmq]"
gRPC server/client interceptors	grpc	pip install "obskit[grpc]"
All integrations (db + kafka + rabbitmq + grpc)	integrations	pip install "obskit[integrations]"
Everything above	all	pip install "obskit[all]"

Quick install

Bash

# Core only — structured logging, health checks, config
pip install obskit

# Full install — every extra included
pip install "obskit[all]"

Zero-boilerplate setup

Python

from obskit import configure_observability, instrument_fastapi
from fastapi import FastAPI

app = FastAPI()

# One call at startup — configures tracing, metrics, logging
obs = configure_observability(
    service_name="order-service",
    environment="production",
    otlp_endpoint="http://tempo:4317",
    trace_sample_rate=0.1,
)

# One call to instrument your framework
instrument_fastapi(app)

log = obs.logger
log.info("order_placed", order_id="123", user_id="u-456")
# → {"event": "order_placed", "order_id": "123", "trace_id": "4bf92f...", "span_id": "00f067..."}

You can also use the individual APIs directly:

Python

from obskit import get_logger, configure_observability

configure_observability(service_name="my-service")
log = get_logger(__name__)
log.info("started")

Why obskit?

• Single package — one pip install obskit, add extras only for what you use
• One setup call — configure_observability() sets up tracing, metrics, and logging in one shot
• Zero import changes — from obskit.logging import get_logger works regardless of which extras are installed
• Graceful degradation — optional integrations (OTel, Prometheus) no-op when their extra is not installed
• 100% test coverage on all components (4,178 tests, enforced in CI)
• PEP 561 typed — full mypy support out of the box