snitchbot 1.2.0

Telegram telemetry: crash reports, slow calls, watchdog and anomaly alerts for Python services

snitchbot

Telegram telemetry for Python services. Crash reports, slow-call alerts, watchdog events, custom notifications, and anomaly detection — delivered to your Telegram chat.

Docs: https://sense1tapo4ek.github.io/snitchbot/

Features

• Crash reporting — uncaught exceptions in sync, threading, and asyncio
• Slow-call monitoring — @watch_slow(threshold_ms=1000) decorator
• Watchdog — event loop block detection with severity escalation
• Custom notifications — notify("deploy complete", severity="warning")
• Request context — trace IDs and tags propagated across async boundaries
• Anomaly detection — RSS spikes, CPU sustained, FD leaks, thread growth
• Interactive Telegram — /status, /last, /test, /mute, /unmute commands
• Live dashboard — auto-updating message with service vitals
• Secret scrubbing — API keys, tokens, passwords never reach Telegram
• Logging bridges — stdlib logging handler + structlog processor

Architecture

Thin host client + detached sidecar process. The client never blocks your app — events are sent via AF_UNIX SOCK_DGRAM to the sidecar, which handles dedup, rate limiting, and Telegram delivery.

Installation

uv add snitchbot

With example frameworks (FastAPI, Litestar, Flask):

uv add snitchbot[examples]

Quick Start

1. Get a Telegram bot token

Talk to @BotFather, create a bot, copy the token.

2. Get your chat ID

Add the bot to a group or start a DM, then use @userinfobot or the Telegram API to find the chat ID.

3. Configure

cp .env.example .env
# Edit .env — set SNITCHBOT_TOKEN and SNITCHBOT_CHAT_ID

4. Add to your app

import snitchbot

# Reads SNITCHBOT_TOKEN and SNITCHBOT_CHAT_ID from env / .env file
snitchbot.init("my-service")

# Or pass explicitly:
# snitchbot.init("my-service", token="...", chat_id="...")

# That's it. Crashes are now reported automatically.

# Optional: custom notifications
snitchbot.notify("Deploy complete", severity="warning")

# Optional: slow-call monitoring
@snitchbot.watch_slow(threshold_ms=1000)
async def process_order(order_id: str):
    ...

# Optional: request context
with snitchbot.request_context(trace_id="req-abc", user_id="42"):
    snitchbot.notify("Processing started")

Running Examples

Prerequisites

# Clone and install
git clone https://github.com/Sense1Tapo4ek/snitchbot.git
cd snitchbot
uv sync --extra dev --extra examples

# Configure
cp .env.example .env
# Edit .env with your bot token and chat ID

FastAPI

uv run uvicorn examples.fastapi_app:app --reload

Then visit:

• http://localhost:8000/ — health check
• http://localhost:8000/notify — sends a custom notification
• http://localhost:8000/slow — triggers slow-call alert (200ms > 100ms threshold)
• http://localhost:8000/crash — triggers crash report
• http://localhost:8000/context — demonstrates request context

Litestar

uv run litestar run --app examples.litestar_app:app --reload

Flask

uv run flask --app examples.flask_app run --reload

Testing

# Run all tests (no Telegram token required)
uv run pytest

# Run only unit tests
uv run pytest tests/shared/unit tests/client/unit tests/sidecar/unit

# Run framework integration tests
uv run pytest tests/e2e/test_framework_integration.py -v

Python Support

• Python >= 3.10
• Linux (primary), macOS (supported)
• Works with: FastAPI, Litestar, Flask, Django, Celery, Gunicorn, uvicorn, uvloop

License

MIT

---