grelmicro 0.41.0

Async-first toolkit. Microservice patterns inside.

A Python toolkit for distributed systems: microservices, modular monoliths, and self-contained systems.

Documentation: https://grelmicro.grel.info/

Source Code: https://github.com/grelinfo/grelmicro

Why grelmicro

grelmicro is an async Python toolkit for microservices and distributed systems: shared locks, caching, rate limits, circuit breakers, retries, and scheduled tasks.

It ships them as small, composable modules with pluggable backends, alongside idempotency keys, the transactional outbox, logging, health checks, metrics, and tracing. Async-first, type-safe, and fully tested.

It is built for any Python application that coordinates work across processes, workers, or replicas. The same primitives serve microservices, a modular monolith, or a self-contained system, and fit naturally into containerized and Kubernetes deployments.

• Micro: one focused primitive per module. Each covers a microservice pattern (distributed lock, leader election, rate limiter, circuit breaker, health check API, externalised configuration).
• Fast: small footprint by design. We keep the layers thin so your code stays quick.
• Async-first: every I/O call is async / await. Drops into FastAPI, FastStream, and any asyncio-based stack.
• Backend-agnostic: each primitive is a protocol. Swap Redis for PostgreSQL or SQLite without touching application code.
• Framework-agnostic: every pattern behaves the same on FastAPI, Starlette, Litestar, and FastStream. micro.install(app) handles the difference, and a parity test holds the claim.
• Railguarded: fully tested, type-checked, and validated. Pre-1.0 the API may change on a minor release. 1.x follows standard semver.

grelmicro is not a task queue (reach for Celery, Dramatiq, or taskiq), not a message broker client (reach for FastStream to publish and subscribe over Kafka, RabbitMQ, NATS, or Redis), and not a web framework (it plugs into FastAPI, Starlette, Litestar, and FastStream). It fills the gap between the web framework you picked and the infrastructure you run.

Already using aiocache, slowapi, pybreaker, tenacity, or aioredlock? See the comparison page for a per-domain breakdown.

Modules

Installation

pip install grelmicro

See the Installation guide for uv and poetry commands, plus optional extras for Redis, PostgreSQL, SQLite, Kubernetes, OpenTelemetry, and structlog.

Example

Run the demo

Want to see every Pattern running against real Redis and Postgres? The FastAPI demo starts in three commands:

cd examples/fastapi-demo
docker compose up --wait
open http://localhost:8000/docs

It wires a cached endpoint, a rate-limited endpoint, a circuit-breaker-protected endpoint, a distributed lock, a leader-gated task, and /healthz / /readyz probes. Read app.py to see each one.

One route, one primitive

The smallest grelmicro program: a FastAPI route protected by a process-local rate limiter. No Grelmicro(...), no Redis, no lifespan.

from fastapi import FastAPI

from grelmicro.providers.memory import MemoryProvider
from grelmicro.resilience import RateLimitExceededError, RateLimiter

app = FastAPI()
api_limiter = RateLimiter.sliding_window(
    "api", limit=100, window=60, backend=MemoryProvider().ratelimiter()
)


@app.get("/ping")
async def ping() -> str:
    try:
        await api_limiter.acquire_or_raise()
    except RateLimitExceededError:
        return "throttled"
    return "ok"

That is the whole thing. Pick a primitive, name it, give it a backend, call it. The memory backend says per-process on purpose. Make it fleet-wide when you need to.

FastAPI with one provider

To make the rate limiter fleet-wide, put one provider in a Grelmicro container and install it into FastAPI. The provider wires a component for every kind it serves, so there is nothing else to list.

from fastapi import FastAPI

from grelmicro import Grelmicro
from grelmicro.providers.redis import RedisProvider
from grelmicro.resilience import RateLimitExceededError, RateLimiter

redis = RedisProvider("redis://localhost:6379/0")
micro = Grelmicro(uses=[redis])

api_limiter = RateLimiter.sliding_window("api", limit=100, window=60)

app = FastAPI()
micro.install(app)


@app.get("/ping")
async def ping() -> str:
    try:
        await api_limiter.acquire_or_raise()
    except RateLimitExceededError:
        return "throttled"
    return "ok"

Adding more primitives is the same shape: they resolve through the same provider. micro.install(app) opens the app on startup, closes it on shutdown, and lets request handlers resolve backends without passing backend=.

FastAPI integration

Create a file main.py with:

import logging
from contextlib import asynccontextmanager

from fastapi import FastAPI, HTTPException, Request
from pydantic import BaseModel

from grelmicro import Grelmicro
from grelmicro.cache import TTLCache, cached
from grelmicro.security import TrustedProxies, resolve_client_address
from grelmicro.health import HealthChecks
from grelmicro.log import configure as configure_logging
from grelmicro.providers.redis import RedisProvider
from grelmicro.resilience import (
    CircuitBreaker,
    RateLimitExceededError,
    RateLimiter,
)
from grelmicro.coordination import LeaderElection, Lock, TaskLock
from grelmicro.task import Tasks

logger = logging.getLogger(__name__)

# === grelmicro app: one container, one lifespan ===
tasks = Tasks()
health = HealthChecks()

# One line says where the shared state lives.
redis = RedisProvider("redis://localhost:6379/0")

leader = LeaderElection("leader-election")
tasks.add_task(leader)

micro = Grelmicro(uses=[redis, tasks, health])

class User(BaseModel):
    id: int
    name: str


# === Patterns declared once at module load, no backend wiring ===
ttl_cache = TTLCache[User](ttl=300)
lock = Lock("shared-resource")
cb = CircuitBreaker("my-service")
api_limiter = RateLimiter.sliding_window("api", limit=100, window=60)


# === FastAPI ===
@asynccontextmanager
async def lifespan(app):
    configure_logging()
    yield


app = FastAPI(lifespan=lifespan)
micro.install(app)


# --- Cache: avoid redundant database queries ---
@cached(ttl_cache)
async def get_user(user_id: int) -> User:
    return User(id=user_id, name="Alice")


@app.get("/users/{user_id}")
async def read_user(user_id: int) -> User:
    return await get_user(user_id)


# --- Circuit Breaker: protect calls to an unreliable service ---
@app.get("/")
async def read_root() -> str:
    async with cb:
        return "Hello World"


# --- Rate Limiter: protect endpoints from overload ---
# Behind a proxy, `request.client.host` is the proxy, so every caller would
# share one bucket. Resolve the real client instead, trusting only your own
# proxies. Drop the `trusted`/`client_key` lines if nothing fronts the app.
trusted = TrustedProxies(["10.0.0.0/8"])


def client_key(request: Request) -> str:
    client = resolve_client_address(request.scope, trusted)
    return client.key if client else "unknown"


@app.get("/api")
async def api_endpoint(request: Request) -> str:
    try:
        await api_limiter.acquire_or_raise(key=client_key(request))
    except RateLimitExceededError as exc:
        raise HTTPException(
            status_code=429,
            detail="Too many requests",
            headers={"Retry-After": str(int(exc.retry_after))},
        )
    return "ok"


# --- Distributed Lock: synchronize access to a shared resource ---
@app.get("/protected")
async def protected() -> str:
    async with lock:
        return "ok"


# --- Interval Task: run locally on every worker ---
@tasks.every(seconds=5)
def heartbeat():
    logger.info("heartbeat")


# --- Distributed Task: run once per interval across all workers ---
@tasks.every(seconds=60, lock=TaskLock(lease_duration=300))
def cleanup():
    logger.info("cleanup")


# --- Leader-gated Task: only the leader executes ---
@tasks.every(seconds=10, leader=leader)
def leader_only_task():
    logger.info("leader task")

The key shape:

• One container, one lifespan. Grelmicro(uses=[...]) lists every Component and active manager. async with micro: opens them all in order, closes in reverse.
• One Provider, many Components. Grelmicro(uses=[redis]) registers a default Component for every kind the RedisProvider serves, and they all share its pool. Name a Component only to override one kind: Grelmicro(uses=[redis, Cache(postgres)]) keeps the rest on Redis.
• Patterns are declared at module load. Lock("cart"), TTLCache(ttl=60), CircuitBreaker("svc") carry no backend reference. They resolve through the active app inside async with, and GrelmicroMiddleware extends that scope to request handlers. The same Lock works in production with Redis and in tests with MemoryLockAdapter, no rewiring.
• Pay only for what you import. import grelmicro does not pull in redis, psycopg, or any other vendor SDK. First-party Providers live under grelmicro.providers.{vendor} and load only when you import them.

For multiple Redis instances, separate names, or test overrides, see the docs.

Contributing

Report bugs and request features in GitHub issues. The reporting guide lists what to include and what happens after you file.

To contribute code or docs, read the contributing guide. It explains the pull request process and the requirements for acceptable contributions: the development setup, the code style, and the pre-merge checklist.

Report security issues privately through the security policy, not a public issue.

License

This project is licensed under the terms of the MIT license.