kubewatch-python

A kubewatch-like application written in Python with Slack support. Can be deployed with the Helm chart in the charts/ subfolder.

Features

• Python asyncio
• Slack "ChatOps" alerts
• Prometheus Metrics endpoint
• Helm Chart

Requirements

• Python >= 3.7

pip

• asyncio
• requests
• kubernetes_asyncio
• prometheus_client

ENVVARs

• SLACK_WEBHOOK (Slack Webook URL)
• CLUSTER (cluster name to show in Slack messages)
• Custom Slack emoji :k8s: (Kubernetes icon)

You can configure the below resources via envvars with the following convention:

• Prefixed with ENABLE_
• Plural form of K8s resource
• Value: true or false

Example

ENABLE_DAEMONSETS or ENABLE_INGRESSES

Watchable Resources

• ClusterRole
• ClusterRoleBinding
• ConfigMap (namespaced, defaults to default namespace)
• CronJob
• DaemonSet
• Deployment
• Ingress
• Job
• LimitRange
• NetworkPolicy
• Node
• PersistentVolume
• PersistentVolumeClaim
• Pod
• PodDisruptionBudget
• PodTemplate
• PodSecurityPolicy
• ReplicaSet
• ResourceQuota
• Role
• RoleBinding
• Secret
• Service
• ServiceAccount
• StatefulSet

Prometheus

The Prometheus metrics endpoint is available at Port 8000.

In addition to the basic Python metrics, there are these two which can be scraped:

• success_webhooks = Counter of how many times the Slack Webhook was successfully triggered
• failed_webhooks = Counter of how many times the Slack Webhook failed

Notes

The first run will give you some history of events