addons.md

Optional Addons for a kubernetes cluster deployment

The configuration variable addons can be used to enable dedicated preconfigured addon deployments provided by the template project. Just add the desired addon names to a map addons in the terraform.tfvarsfile in the cluster project.

  addons = {
    "dashboard" = { }
    "nginx-ingress" = { }
    "heapster" = { }
  }

The value of every entry must be the local addon config as map. If there are no special settings the map should be empty. It will automatically be backed by default config for the dedicated addon.

By default the addons dashboard, nginx and heapster are enabled. This can be changed by explicitly setting the variable addons in the cluster configuration and omitting the dedocazted entry.

The following addons are supported so far:

Name	Description
dashboard	The standard kubernetes dashboard
heapster	Resource usage analysis
nginx-ingress	Nginx http/https ingress controller for ports 80 and 443
kube-lego	Letsencrypt based TLS certificate support
logging	Fluentd, Kibana, ElasticSearch logging stack
monitoring	Prometheus, Alertmanager, Grafana monitoring stack
guestbook	Application Deployment Example
dex	Identity Provider
gardener	Kubernetes Cluster Management

Monitoring Addon – Add custom config

The monitoring stack, which consists of Prometheus, Alertmanager and Grafana, can be extended by injection of custom configuration. The custom configuration, provided in static files, will be appended and/or merged to the default configuration.

An exception here is the alertmanager. The config of those component will be overridden by custom config. Only some global configuration for smtp will be not overridden.

Config snippet to add to the terraform.tfvars. Everthing is optional (smtp config makes no sense if not provided completly).

addons = {
  "monitoring" = {
    "prometheus_rules_file"    = "<relative/path/to/custom/prometheus/rules.yaml>"
    "prometheus_config_file"   = "<relative/path/to/custom/prometheus/scrape/config.yaml>"
    "grafana_config_file"      = "<relative/path/to/custom/grafana/dashboard/config.yaml>"
    "alertmanager_config_file" = "<relative/path/to/custom/alertmanager/config.yaml>"

    "smtp_host"     = "<global-smtp-host>"
    "smtp_from"     = "<global-smtp-from>"
    "smtp_username" = "<global-smtp-user>"
    "smtp_password" = "<global-smtp-password>"
  }
}

Example config

1. Prometheus scrape configs

- job_name: 'test'
  ...
- job_name: 'test-2'
  ...

2. Prometheus rules

alerts-1.yaml: |
  groups:
  - name: test.rules
    rules:
    - alert: test
      expr: ...
      for: 0s
  ...

3. Grafana Dashboard

dashboard-1.json: |-
  {
    dashboard: {
      ...
    }
  }
dashboard-2.json: |-
  {
    dashboard: {
      ...
    }
  }

4. Alertmanager config

route:
  receiver: dev-null
  ...
receivers:
- name: dev-null
  ...