Archived: due to no longer supporting within a deployment and never given any more time to tackle its many issues and bugs.

Kubernetes sidecar installed alongside Grafana to deploy dashboards. Utilises shared volume between sidecar and Grafana containers. Grafana automatically detects new Dashboards placed into configured data source.

Replacement for sidecar included with the Helm Prometheus Monitoring Stack.

• Kubernetes Custom Resource Definition for Grafana Dashboard.
• Utilises the Kopf framework.
• Example Gatekeeper configuration for constraints on dashboard resources.

• local development environment setup for testing and development

Documentation

• Changes introduced through PRs.
  • update changelog, readme and documentation
• Tag against main branch.
  • Semantic Versioning prefixing with a v.
• Include auto generated release notes curating where necessary.

virtualenv venv
source venv/bin/activate             # activate the venv

Install dependencies defined in setup.py.

pip install --editable .[dev]        # install dependencies including dev dependencies

Create testable executable: grafana-k8-sidecar

Following documentation makes use of kind for a local K8 cluster to allow for testing.

Kind Images

kind create cluster --wait 5m     # specific version: `--image=kindest/node:<VERSION>`
export KUBE_CTX=kind-kind         # set kubectl to use kind cluster
kind delete cluster               # `--name` to delete specific cluster

Install latest Gatekeeper release:

kubectl apply -f https://raw.githubusercontent.com/open-policy-agent/gatekeeper/<VERSION>/deploy/gatekeeper.yaml
kubectl apply -f https://raw.githubusercontent.com/open-policy-agent/gatekeeper/<VERSION>/demo/basic/sync.yaml

Run Terraform deployment.

cd ./terraform
terraform init
terraform apply

Review Grafana Dashboard resources created as part of Terraform deployment.

kubectl get grafanadashboards --all-namespaces

source venv/bin/activate  # activate the venv
grafana-k8-sidecar        # this will use defaults (max-workers: 20, working-dir: /tmp/grafana-dashboards)

Settings:

• --working-dir=./sidecar/tests/fixtures/dashboards to the dashboard fixtures.
• --max-workers=1 max workers to 1 for easier chronological debugging.

Sidecar exposes prometheus metrics.

• unit
  • pytest
  • configured to run as part of pre-commit config.
• system
  • pytest -W ignore
  • ToDo: currently broken and set to ignore.
• fixtures

Local monitoring setup for testing metrics created by service.

cd local                    # dir with docker-compose and config files
docker-compose up           # bring up the containers
docker-compose down -v      # tear down containers and remove volumes

Available services:

• grafana
  • username: admin,
  • password: test
• prometheus