console.txt

# HELP deployment_status Status of deployments
# TYPE deployment_status gauge
deployment_status{deployment="lib",namespace="arc"} 1.0
deployment_status{deployment="studio",namespace="arc"} 1.0
deployment_status{deployment="harmony",namespace="arc"} 1.0
deployment_status{deployment="prose",namespace="arc"} 1.0
deployment_status{deployment="nginx-deployment",namespace="arc"} 1.0
deployment_status{deployment="studio-dashboard",namespace="arc"} 1.0
deployment_status{deployment="melody",namespace="arc"} 1.0
# HELP available_cpu Available CPU in the namespace
# TYPE available_cpu gauge
available_cpu{namespace="arc"} +Inf
# HELP available_memory Available memory in the namespace
# TYPE available_memory gauge
available_memory{namespace="arc"} +Inf   

ns.py

import os
import time
from kubernetes import client, config
from prometheus_client import start_http_server, Gauge

NAMESPACE = os.getenv("NAMESPACE", "default")  # Get namespace from environment variable or default to "default"
KUBE_CONFIG_PATH = os.getenv("KUBE_CONFIG_PATH", "~/.kube/config")


def get_kubernetes_client():
    try:
        # Use kube config for authentication
        config.load_kube_config(config_file=KUBE_CONFIG_PATH)
        return client.ApiClient()
    except Exception as e:
        print("Failed to load kube config:", e)
        return None


def main():
    k8s_client = get_kubernetes_client()
    if k8s_client is None:
        print("Exiting...")
        return

    v1 = client.AppsV1Api(k8s_client)

    # Prometheus metrics
    deployment_status = Gauge('deployment_status', 'Status of deployments', ['deployment', 'namespace'])

    while True:
        deployments = v1.list_namespaced_deployment(namespace=NAMESPACE).items

        for deployment in deployments:
            deployment_name = deployment.metadata.name
            desired_replicas = deployment.spec.replicas

            # Get current number of pods for the deployment
            current_replicas = deployment.status.replicas if deployment.status.replicas else 0

            # Check if the deployment is up or down
            status = 1 if current_replicas == desired_replicas else 0

            # Set Prometheus metric
            deployment_status.labels(deployment=deployment_name, namespace=NAMESPACE).set(status)

        # Sleep for a minute before checking again
        time.sleep(60)


if __name__ == '__main__':
    start_http_server(8000)  # Start Prometheus client on port 8000
    main()