kubewatch

The Kubewatch Controller initializes its configuration using a specific hierarchy of sources:

1. Configuration File: It first attempts to read a .kubewatch.yaml file.
2. Command Line Arguments: Arguments passed during execution.
3. Environment Variables: If parameters are missing from the YAML file or CLI arguments, the controller falls back to reading standard environment variables.

The resulting configuration object determines which resources are watched and which handlers are activated.

The ValidateValue mechanism can be used to ensure specific configuration values are not empty. If a required value is missing or empty, Helm will throw an error during installation or upgrade, providing instructions on how to retrieve the missing value from a secret.

Example error behavior: If a value is required but provided as an empty string via --set:

$ helm install test mychart --set path.to.value00=""
    'path.to.value00' must not be empty, please add '--set path.to.value00=$PASSWORD_00' to the command. To get the current value:

        export PASSWORD_00=$(kubectl get secret --namespace default secretName -o jsonpath="{.data.password-00}" | base64 --decode)

Kubewatch operates using three core components that manage the lifecycle of a Kubernetes event from detection to notification:

1. Config: Defines the operational parameters, including which handlers to use and which filters to apply. This object is used to initialize the client.
2. Controller: The engine that watches Kubernetes resources. It uses SharedIndexInformers to listen for resource changes via the Kubernetes API Server. When events occur, the Controller places them in a rate-limiting queue, filters them based on your configuration, and then dispatches them to the appropriate Handler.
3. Handler: The execution layer. Once an event is filtered, the Handler is responsible for processing it and sending the notification to the configured destination (e.g., Slack, Email, etc.).

If you have existing Kubernetes secrets containing sensitive data (like passwords), you can use the ExistingSecret schema to map those existing keys to the expected keys in your deployment.

Key fields:

• name: The name of the existing Kubernetes secret.
• keyMapping: An object that maps the expected key name in your application to the actual key name inside the existing secret.

Example usage in a deployment template:

# values.yaml
name: mySecret
keyMapping:
  password: myPasswordKey

# templates/dpl.yaml (logic snippet)
      env:
        - name: PASSWORD
          valueFrom:
            secretKeyRef:
              name: {{ include "common.secrets.name" (dict "existingSecret" .Values.existingSecret "context" $) }}
              key: {{ include "common.secrets.key" (dict "existingSecret" .Values.existingSecret "key" "password") }}

To use the common logic and template helpers provided by the Bitnami Common Library Chart in your own Helm chart, add it to your Chart.yaml under the dependencies section. After adding the dependency, run helm dependency update to download the library chart.

dependencies:
  - name: common
    version: 1.x.x
    repository: https://charts.bitnami.com/bitnami

$ helm dependency update

For a quick installation using kubectl, you can use a ConfigMap to hold your configuration and a Pod/Deployment to run the watcher.

1. Create the ConfigMap (ensure you update the Slack channel and token):

kubectl create -f kubewatch-configmap.yaml

2. Create the Pod or Deployment:

kubectl create -f kubewatch.yaml

Note: A kubewatch container will be created along with a kubectl sidecar container to allow communication with the API server.

To remove all Kubernetes components associated with the Kubewatch deployment and delete the Helm release, use the helm delete command with your release name.

$ helm delete my-release

Upgrading to version 3.0.0 involves significant changes:

• Chart labels now follow standard Helm chart practices.
• It introduces bitnami/common as a library chart dependency.

Important: Backwards compatibility is not guaranteed. To upgrade to 3.0.0, you should install a new release of the Kubewatch chart rather than performing an in-place upgrade. Ensure you update chart dependencies before executing the upgrade.

To remove the Kubewatch installation, use the standard Helm uninstall command with the name of your release.

helm uninstall <release-name>