Skip to main content

loopmon - Lightweight Event Loop monitoring

Codecov Dependabot Status PyPI PyPI - Wheel Python Version GitHub last commit (branch) GitHub Workflow Status (branch) Code Style

loopmon is a lightweight library that can detect throttling of event loops. For example, you can detect long-running coroutine or whether the blocking function is invoked inside the event loop.

Usage

import asyncio
import time
from datetime import datetime
import loopmon

async def print_collected_data(lag: float, tasks: int, data_at: datetime) -> None:
    print(f'event loop lag: {lag:.3f}, running tasks: {tasks}, at {data_at}')

async def main() -> None:
    loopmon.create(interval=0.5, callbacks=[print_collected_data])
    # Simple I/O bound coroutine does not occur event loop lag
    await asyncio.sleep(0.2)
    # Blocking function call
    time.sleep(1)

if __name__ == '__main__':
    asyncio.run(main())

will prints:

event loop lag: 0.000, running tasks: 2, at 2022-02-24 13:29:05.367330+00:00
event loop lag: 1.001, running tasks: 1, at 2022-02-24 13:29:06.468622+00:00

You can check other examples.

I recommend you to add loopmon.create(...) on beginning of async function if you are not familiar with handling loop itself. But you can also control creation, installation or staring of monitor via EventLoopMonitor.start() or EventLoopMonitor.install_to_loop().

Features

  • Detects event loop lag
    • Detects event loop running on other thread. example
  • Collect how many tasks are running in the event loop
  • Customize monitoring start and end points
  • Customize monitoring interval
  • Customize collected metrics through callbacks
  • 100% type annotated
  • Zero dependency (except typing-extentions)

How it works

Event loop is single threaded and based on Cooperative Scheduling. So if there is a task that does not yield to another tasks, any tasks on the loop can not be executed. And starvation also happens when there are too many tasks that a event loop can not handle.

Currently loopmon.SleepEventLoopMonitor is one and only monitor implementation. It periodically sleeps with remembering time just before sleep, and compares the time after awake. The starvation happen if the difference bigger than its sleeping interval.

pseudo code of SleepEventLoopMonitor

while True:
    before = loop.time()
    await asyncio.sleep(interval)
    lag = loop.time() - before - interval
    tasks = len(asyncio.all_tasks(loop))
    data_at = datetime.now(timezone.utc)
    for c in callbacks:
        loop.create_task(c(lag, tasks, data_at))

Integration examples

Prometheus

from datetime import datetime
from functools import partial
import loopmon
from prometheus_client import Gauge

async def collect_lag(gauge: Gauge, lag: float, _: int, __: datetime) -> None:
    gauge.set(lag)

async def main(gauge: Gauge) -> None:
    loopmon.create(interval=0.5, callbacks=[partial(collect_lag, gauge)])
    ...

Metadata

Release files for loopmon 1.0.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for loopmon 1.0.1
File Size Uploaded
loopmon-1.0.1.tar.gz 6.7 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for loopmon 1.0.1
File Interpreter ABI Platform
loopmon-1.0.1-py3-none-any.whl Python 3 none any Details

Total release size: 14.0 kB

Release files / loopmon-1.0.1.tar.gz

Download URL loopmon-1.0.1.tar.gz
Size 6.7 kB
Tags Source
SHA-256 checksum
How to use checksums
c259dd82b14c05ac2007eb3a4cb4e4d5f79cac931ce89556e9131ae5dc1abb4c
BLAKE2b-256 checksum
How to use checksums
2f359a8dc9f6ae01bb7d5f543b4da31626845687a4587f817f0507423a8e5756
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/4.0.1 CPython/3.11.3

Release files / loopmon-1.0.1-py3-none-any.whl

Download URL loopmon-1.0.1-py3-none-any.whl
Size 7.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
96ef43eaa3763804c7bc2f4aaf849a54146192c25da2680cd8ebc13856a2d37f
BLAKE2b-256 checksum
How to use checksums
a51001f09de2eff16dac02a2cba91671fa5612944fa38a81a3da4470510696eb
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/4.0.1 CPython/3.11.3

Release history Release notifications | RSS feed

This release

1.0.1 This release

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page