Query cost monitoring API with Grafana

You can directly query the cost monitoring API using the same Grafana token provisioned in Programmatically accessing Prometheus data through Grafana.

Note

The cost monitoring API returns data as JSON blobs, unlike standard Prometheus time series data. This means that you cannot configure the cost monitoring API as a standard Prometheus datasource for services such as in-built AWS CloudWatch Dashboards.

Prerequisites

• A Grafana service account token (see Programmatically accessing Prometheus data through Grafana for instructions)

• A .env file to securely store your Grafana token (see Secrets, passwords and access tokens)

• A virtual environment with the following Python packages installed:

  • python-dotenv

  • requests

  • yarl

Sample Code

import json
import os
import requests
from dotenv import load_dotenv
from yarl import URL

# Load Grafana service account token from .env file
load_dotenv()
grafana_token = os.environ["GRAFANA_TOKEN"]

grafana_url = URL("https://grafana.<cluster_name>.2i2c.cloud/")
api_url = grafana_url / "api/datasources"
datasource = requests.get(
    api_url,
    headers={
        "Accept": "application/json",
        "Content-Type": "application/json",
        "Authorization": f"Bearer {grafana_token}",
    }
)

# Get the cost-monitoring yesoreyeram-infinity-datasource UID
datasource_uid = datasource.json()[1]["uid"]
target_url = URL("http://jupyterhub-cost-monitoring.support.svc.cluster.local")
# See https://jupyterhub-cost-monitoring.readthedocs.io/en/latest/api/ for available endpoints
subpath = "total-costs"

headers = {
    "Authorization": f"Bearer {grafana_token}",
    "Content-Type": "application/json"
}

payload = {
    "queries": [
        {
            "datasource": {
                "type": "yesoreyeram-infinity-datasource",
                "uid": datasource_uid,
            },
            "url": str(target_url / subpath),
            "url_options": {
                "method": 'GET',
                "data": '',
            },
            "type": "json",
            "source": "url",
            "format": "table",
            "parser": "backend",
            "root_selector": "",
            # Columns can change depending on data returned by endpoint – see https://github.com/2i2c-org/jupyterhub-cost-monitoring/blob/a6e49c719e8600fc2490e2fb2a77da09ce4bcd1a/dashboards/common.libsonnet#L128 for examples from existing cloud cost grafana dashboards
            "columns": [
              {"selector": "date", "text": "Date", "type": "timestamp"},
              {"selector": "name", "text": "Name", "type": "string"},
              {"selector": "cost", "text": "Cost", "type": "number"},
            ],
        }
    ],
    "from": "now-1d",
    "to": "now"
}

# https://grafana.com/docs/grafana-cloud/developer-resources/api-reference/http-api/data_source/#query-a-data-source
query_url = grafana_url / "api/ds/query"

response = requests.post(
    query_url,
    headers=headers,
    data=json.dumps(payload)
)

response.raise_for_status()
data = response.json()

print(json.dumps(data, indent=2))