tqdm-prometheus-exporter 0.1.14

Echo tqdm progress to a prometheus endpoint

Prometheus exporter for tqdm

Publish your pretty CLI stats for monitoring elsewhere

Usage

There are no security or authentication mechanisms provided, be careful exposing the host more widely

Initalise the proxy

proxy = TqdmPromProxy(http_host='[::1]', http_port=3000, metric_func=metric_formatter_func)
proxy.start()

Use the proxy

Replace tqdm initalisers with proxy.tqdm.

for item in proxy.tqdm(range(100), desc='My first bar')
    item.doathing()

with proxy.tqdm(range(100), desc='My second bar') as f:
    f.update(50)

# etc

Formatting your content

TQDM bars are a lot more verbose than prometheus metrics. You can customise their names by passing in a function with this signature.

def metric_formatter_func(metric_name: str, metric_property: str, metric_scale: str) -> str:
    pass

metric_property is one of:

• completed the number executions thus far in TQDM;
• total the configured expected maximum in TQDM as tqdm(total=x);
• active based on the naming, how many current bars are being grouped together as this name
• finished based on the naming, how many bars with this name have been completely finished

The str that is returned must adhere to the metric naming conventions (A-Za-z0-9_) outlined in the prometheus docs https://prometheus.io/docs/practices/naming/

This function is called a lot, so avoid computation or cache your values

Additionally to control which properties are aggregated, provide a bucket_func: This example will aggregate all bars with a description starting with upload together

def bucketting_func(snapshot: TqdmSnapshot):
    if snapshot.desc.lowercase().startswith('upload'):
        return "Upload"

    return snapshot.desc

If you require tqdm.write for sticky console output,

from tqdm import tqdm as tqdm_og
tqdm_og.write('Job done')

curl http://[::1]:3000/metrics

[!WARNING]
For windows users ensure you add this to your __main__ block, and before any initialisation of TqdmPrometheusProxy

from multiprocessing import freeze_support
freeze_support()

Overview

---
Conceptual Overview
---
flowchart LR

    Queue@{ shape: das, label: "Update buffer" }
    TQDM@{ shape: process, label: "(your) faux TQDM instance"}
    Bucketting@{ shape: process, label: "Metric extraction & aggregation"}

    TQDM -- "Periodic Updates" --> Queue
    Queue --> Bucketting
    Bucketting -- "Expose aggregate metrics" --> HttpServer