We received many requests from PythonAnywhere users to make it possible to programmatically monitor usage of CPU credit, so we decided to add a new endpoint to our experimental API.
The first step when using the API is to get an API token – this is what you use to authenticate yourself with our servers when using it. To do that, log in to PythonAnywhere, and go to the “Account” page using the link at the top right. Click on the “API token” tab; if you don’t already have a token, it will look like this:
Click the “Create a new API token” button to get your token, and you’ll see this:
That string of letters and numbers (d870f0cac74964b27db563aeda9e418565a0d60d in the screenshot) is an API token, and anyone who has it can access your PythonAnywhere account and do stuff – so keep it secret. If someone does somehow get hold of it, you can revoke it on this page by clicking the red button – that stops it from working in the future, and creates a new one for you to use.
Now you can use CPU API to track your CPU usage.
For example you could use it at the beginning and the end of your script to learn how many CPU seconds were consumed (assuming nothing else is running).
from math import factorial
from time import sleep
from urllib.parse import urljoin
import requests
api_token = "YOUR TOKEN HERE"
username = "YOUR USERNAME HERE"
pythonanywhere_host = "www.pythonanywhere.com"
api_base = "https://{pythonanywhere_host}/api/v0/user/{username}/".format(
pythonanywhere_host=pythonanywhere_host,
username=username,
)
resp = requests.get(
urljoin(api_base, "cpu/"),
headers={"Authorization": "Token {api_token}".format(api_token=api_token)}
)
initial_usage_seconds = resp.json()["daily_cpu_total_usage_seconds"]
# we burn some cpu seconds
[factorial(x) for x in range(2000)]
# cpu usage is updated every 60 seconds, so we need to wait to be sure that usage is available for us to read.
sleep(70)
resp = requests.get(
urljoin(api_base, "cpu/"),
headers={"Authorization": "Token {api_token}".format(api_token=api_token)}
)
final_usage_seconds = resp.json()["daily_cpu_total_usage_seconds"]
seconds_used = final_usage_seconds - initial_usage_seconds
print("Task cost {} CPU seconds to run".format(seconds_used))
…replacing "YOUR TOKEN HERE" and "YOUR USERNAME HERE" with the appropriate stuff.
If you’re on our EU-based system, you should also replace www.pythonanywhere.com with
eu.pythonanywhere.com.
Let us know if you have any comments or questions – otherwise, happy coding!