Try Pytheus multiprocess without changing your prometheus code!

Shy Devy — Sun, 08 Oct 2023 11:01:01 +0000

There is an extremely experimental feature that allows you to patch in-place the official prometheus_client library.

If you have issues with the official client or just want to try it out it's as simple as:

import prometheus_client
from pytheus.experimental.compatibility import patch_client

patch_client(prometheus_client)

This has to be done at the start/initialization of your application and you can load a different backend before applying with patch with patch_client by using load_backend.

This supports all the metrics types Counter, Gauge, Summary & Histogram. It supports the default REGISTRY and generate_latest. It also works with the start_http_server function.

Of course it's extremely limited and you might be using a feature that is not yet supported, but feel free to try! :)

(This was used to test an internal library without having to migrate all of it, so since I worked on it might as well share it! )

https://github.com/Llandy3d/pytheus

[pytheus] simple multiprocess metrics for sync/async python applications

Shy Devy — Sat, 01 Jul 2023 15:20:59 +0000

pytheus

A modern python library for collecting prometheus metrics for your application that works both in synchronous and asyncio programs.

docs: https://pythe.us

Llandy3d / pytheus

experimenting with a new prometheus client for python

pytheus

playing with metrics

Documentation: https://pythe.us

pytheus is a modern python library for collecting prometheus metrics built with multiprocessing in mind.

Some of the features are:

multiple multiprocess support
- redis backend ✅
- Rust powered backend 🧪
- bring your own ✅
support for default labels value ✅
partial labels value (built in an incremental way) ✅
customizable registry support ✅
registry prefix support ✅

Philosophy

Simply put is to let you work with metrics the way you want.

Be extremely flexible, allow for customization from the user for anything they might want to do without having to resort to hacks and most importantly offer the same api for single & multi process scenarios, the switch should be as easy as loading a different backend without having to change anything in the code.

What you see is what you get.
No differences between singleprocess & multiprocess, the only change is…

View on GitHub

Usage

Install with:

pip install pytheus

create a counter and increment it:

from pytheus.metrics import Counter

cache_hit_total = Counter(name='cache_hit_total', description='desc')
cache_hit_total.inc()

finally generate the metrics:

from pytheus.exposition import generate_metrics

metrics = generate_metrics()

Full example with a flask application

import time
from flask import Flask, Response
from pytheus.metrics import Histogram
from pytheus.exposition import generate_metrics, PROMETHEUS_CONTENT_TYPE

app = Flask(__name__)

http_request_duration_seconds = Histogram(
    'http_request_duration_seconds', 'documenting the metric..'
)

@app.route('/metrics')
def metrics():
    data = generate_metrics()
    return Response(data, headers={'Content-Type': PROMETHEUS_CONTENT_TYPE})

# track time with the context manager
@app.route('/')
def home():
    with http_request_duration_seconds.time():
        return 'hello world!'

# alternatively you can also track time with the decorator shortcut
@app.route('/slow')
@http_request_duration_seconds
def slow():
    time.sleep(3)
    return 'hello world! from slow!'

app.run(host='0.0.0.0', port=8080)

Rust powered multiprocess 🦀

For multiprocess in python you need to synchronize the metrics between all the gunicorn workers for example, this backend implementation makes it extremely easy for your application to support that and it's also compatible with asyncio python applications as the processing will be done in parallel under the hood!

Install:

pip install pytheus-backend-rs

Load the backend:

from pytheus.backends import load_backend
from pytheus_backend_rs import RedisBackend

load_backend(
    backend_class=RedisBackend,
    backend_config={"host": "127.0.0.1", "port": 6379},
)

and that's it! Now regardless of how many workers you use the metrics will be in sync between them :)

FastAPI automatic metrics collection

The library offer a middleware for FastAPI applications that will automatically collect metrics regarding http request processing times & size of data, with the possibility of inspecting it depending on method (GET/POST/..), status_code (200,400,..) & route (/home).

Load the middleware

from fastapi import FastAPI
from pytheus.middleware import PytheusMiddlewareASGI


app = FastAPI()
app.add_middleware(PytheusMiddlewareASGI)

Create the endpoint for prometheus to collect metrics

from fastapi.responses import PlainTextResponse
from pytheus.exposition import generate_metrics


@app.get('/metrics', response_class=PlainTextResponse)
def pytheus_metrics():
    return generate_metrics()

If you visit the endpoint you will see metrics like:

# HELP http_request_duration_seconds duration of the http request
# TYPE http_request_duration_seconds histogram
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.005"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.01"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.025"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.05"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.1"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.25"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="0.5"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="1"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="2.5"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="5"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="10"} 1.0
http_request_duration_seconds_bucket{method="GET",route="/metrics",status_code="200",le="+Inf"} 1.0
http_request_duration_seconds_sum{method="GET",route="/metrics",status_code="200"} 0.0014027919969521463
http_request_duration_seconds_count{method="GET",route="/metrics",status_code="200"} 1.0
# HELP http_request_size_bytes http request size
# TYPE http_request_size_bytes histogram
# HELP http_response_size_bytes http response size
# TYPE http_response_size_bytes histogram
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="10.0"} 0.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="100.0"} 0.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="1000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="10000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="100000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="1000000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="10000000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="100000000.0"} 1.0
http_response_size_bytes_bucket{method="GET",route="/metrics",status_code="200",le="+Inf"} 1.0
http_response_size_bytes_sum{method="GET",route="/metrics",status_code="200"} 296.0
http_response_size_bytes_count{method="GET",route="/metrics",status_code="200"} 1.0

DEV Community: Shy Devy

Try Pytheus multiprocess without changing your prometheus code!

[pytheus] simple multiprocess metrics for sync/async python applications

pytheus

Llandy3d / pytheus

experimenting with a new prometheus client for python

pytheus

Philosophy

Usage

Full example with a flask application

Rust powered multiprocess 🦀

FastAPI automatic metrics collection

Load the middleware

Create the endpoint for prometheus to collect metrics