No project description provided
Project description
Prometheus Distributed Client
Purpose and principle
prometheus-distributed-client
is aimed at shorted lived process that can expose Prometheus metrics through HTTP.
Advantages over Pushgateway
The prometheus project provides several ways of publishing metrics. Either you publish them directly like the official client allows you to do, or you push them to a pushgateway.
The first method implies you've got to keep your metrics in-memory and publishs them over http. The second method implies that you'll either have a pushgateway per process or split your metrics over all your processes to avoid overwriting your existing pushed metrics.
prometheus-distributed-client
allows you to have your short lived process push metrics to a database and have another process serving them over HTTP. One of the perks of that approach is that you keep consistency over concurrent calls. (Making multiple counter increment from multiple process will be acknowledge correctly by the database).
Code examples
prometheus-distributed-client
uses the base of the official client but replaces all write and read operation by database call.
Declaring and using metrics
from prometheus-distributed-client import set_redis_conn, Counter, Gauge
# we use the official clients internal architecture
from prometheus_client import CollectorRegistry
# set your own registry
REGISTRY = CollectorRegistry()
# declare metrics from prometheus-distributed-client
COUNTER = Counter('counter_metric_name', 'metric documentation',
[labels], registry=REGISTRY)
GAUGE = Gauge('gauge_metric_name', 'metric documentation',
[labels], registry=REGISTRY)
# increment a counter and set a value for a gauge
COUNTER.labels('label_value').inc()
GAUGE.labels('other_label_value').set(12)
Serving the metrics
prometheus-distributed-client
use the registry system from the official client and is de facto compatible with it. If you want to register regular metrics alongside the one from prometheus-distributed-client
it is totally feasible.
Here is a little example of how to serv metrics from prometheus-distributed-client
, but you can also refer to the documentation of the official client.
# with flask
from flask import Flask
from prometheus_client import generate_latest
# get the registry you declared your metrics in
from metrics import REGISTRY
app = Flask()
@app.route('/metrics')
def metrics():
return generate_latest(REGISTRY)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for prometheus_distributed_client-1.2.2.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | ae26c4f539e6b3e69e6086ff0e1376e82e49c1ad65ccf5eff1fc17e95debc278 |
|
MD5 | 2ff495809984f5b702fceda40f0c8c62 |
|
BLAKE2b-256 | 94e39d75aea666a46820c7b6aad09d65ab45373fe5c9f2a275974a106c3abf63 |
Hashes for prometheus_distributed_client-1.2.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 194fb46cd990d72403936b0d0569b24a36cb69ba8bd951f656d2a7ea5124b4bd |
|
MD5 | 0088f37c658d9bad863410e7d2e4a838 |
|
BLAKE2b-256 | 61af844ab330b77f0805f1d32b422f7c661ce3f96c2a3baf27f8a4419ff83ca6 |