🩺 flask-vitals

A drop-in Flask health blueprint — /healthz (liveness) + /readyz (readiness) in two lines.

---

Every service needs the same two endpoints — "are you alive?" and "can you serve?" — yet everyone re-writes them, slightly differently, in every app. flask-vitals is the boring, correct version you register once:

from flask_vitals import vitals

app.register_blueprint(vitals())

…and now GET /healthz and GET /readyz answer exactly the way orchestrators (Kubernetes), load balancers, and external monitors (Huginn, UptimeRobot, Uptime Kuma) expect.

✨ Features

• Liveness vs. readiness, done right — /healthz says the process is up; /readyz says it can actually serve (its dependencies are reachable).
• Injected checks — readiness probes your DB / cache / queue via callables you provide; the library never reaches into your stack.
• Honest status codes — /readyz returns 503 with a per-check breakdown when something is down, so "up but broken" is caught.
• Zero config to start, fully tunable — custom paths, an optional version string, any number of named checks.
• Tiny & typed — one module, no dependency beyond Flask, mypy --strict, 100 % tested. MIT-licensed.

📦 Install

pip install flask-vitals
# or
uv add flask-vitals

🚀 Quickstart

from flask import Flask
from sqlalchemy import text
from flask_vitals import vitals

app = Flask(__name__)

def db_ok() -> None:
    db.session.execute(text("SELECT 1"))   # raises → not ready

app.register_blueprint(vitals(checks={"db": db_ok}, version="1.4.2"))

Request	Response	Code
GET /healthz	{"status": "ok", "version": "1.4.2"}	200
GET /readyz	{"ready": true, "checks": {"db": true}}	200
GET /readyz	{"ready": false, "checks": {"db": false}} (db unreachable)	503

A check passes unless it returns False or raises — so a one-liner like lambda: cache.ping() is a valid check.

Options

Argument	Default	Purpose
checks	{}	{name: callable} readiness probes
version	None	echoed by /healthz when set
liveness_path	/healthz	liveness route
readiness_path	/readyz	readiness route
name	"vitals"	blueprint name (change if registered twice)

📡 Monitoring it (e.g. with Huginn)

Point an uptime monitor at each app:

• an HTTP monitor on …/healthz — expected status 200 → is it alive?
• a keyword monitor on …/readyz containing "ready": true → is it serving?

That's the UptimeRobot/Uptime-Kuma experience for any Flask app, with no bespoke health code per service.

🛠️ Develop

uv sync
uv run ruff format --check .
uv run ruff check .
uv run mypy
uv run pytest

🧱 Architecture

One module, flask_vitals.blueprint, exposing vitals(...) -> flask.Blueprint. No global state, no dependency beyond Flask; readiness checks are injected by the host app, so the library stays decoupled from your stack and trivially testable.

🤝 Contributing

Issues and PRs are welcome! Keep it small and tested:

1. uv sync
2. Make your change with a test.
3. uv run ruff format . && uv run ruff check . && uv run mypy && uv run pytest
4. Open a PR.

📄 License

MIT © Clara Vanacker