Skip to content

Latest commit

 

History

History
85 lines (66 loc) · 3.26 KB

CONTRIBUTING.md

File metadata and controls

85 lines (66 loc) · 3.26 KB
Raw

Contributing to grafana-k8s

GitHub License GitHub Commit Activity GitHub Lines of Code GitHub Issues GitHub PRs GitHub Contributors GitHub Watchers

Development environment

Initial setup (ensure microk8s is a clean slate with microk8s.reset or a fresh install with snap install microk8s --classic:

microk8s.enable dns storage
juju bootstrap microk8s mk8s
juju add-model lma
juju create-storage-pool operator-storage kubernetes storage-class=microk8s-hostpath

Deploy Grafana on its own:

git clone git@github.com:canonical/grafana-k8s.git
cd grafana-k8s
charmcraft pack
juju deploy ./grafana-k8s_ubuntu-20.04-amd64.charm --resource grafana-image=ubuntu/grafana:latest --resource litestream-image=docker.io/litestream/litestream:0.4.0-beta.2

View the dashboard in a browser:

  1. juju status to check the IP of the running Grafana application
  2. Navigate to http://IP_ADDRESS:3000
  3. Log in with the default credentials username=admin, password=admin.

Add Prometheus as a datasource. See the contributing guide for Prometheus to build and deploy, then:

juju add-relation grafana-k8s prometheus-k8s
watch -c juju status --color  # wait for things to settle down

Once the deployed charm and relation settles, you should be able to see Prometheus data propagating to the Grafana dashboard.

High Availability Grafana

This charm is written to support a high-availability Grafana cluster, but a database relation is required (MySQL or Postgresql).

If HA is not required, there is no need to add a database relation.

NOTE: HA should not be considered for production use.

...

Developing

Create and activate a virtualenv, and install the development requirements,

virtualenv -p python3 venv
source venv/bin/activate
pip install -r requirements-dev.txt

Testing

The tests are run with tox; the following tox targets are available:

  • lint runs the linting checks based on pflake8, isort and black, in order; the first to fail interrupts the tox -e lint run.
  • static runs static checks with mypy.
  • unit runs the unit tests.

Debugging

Data sources

When data sources are related to grafana, they should appear in the datasources.yaml file. This can be manually verified by ssh-ing into the grafana container:

juju ssh --container grafana grafana/0
cat /etc/grafana/provisioning/datasources/datasources.yaml

or querying the grafana HTTP API:

curl --user admin:password http://IP_ADDRESS:3000/api/datasources/