blumeops/argocd/manifests/grafana/values.yaml
Erich Blume 0cdc143227 Deploy Dex OIDC identity provider with Grafana SSO (#222)
## Summary
- Deploys Dex OIDC identity provider on ringtail k3s cluster as central authentication service
- Integrates Grafana as first SSO client via `auth.generic_oauth`
- Uses Kubernetes CRD storage backend (no PVC needed)
- All secrets (bcrypt hash, client secrets) injected via ExternalSecrets from 1Password item "Dex (blumeops)"
- NixOS-built container image via `containers/dex/default.nix`

## Pre-requisites (manual, before deployment)
1. Create 1Password item "Dex (blumeops)" in `blumeops` vault with fields:
   - `password`: strong generated password for Dex login
   - `static-password-hash`: bcrypt hash of above (`htpasswd -BnC 10 eblume`, copy hash after `eblume:`)
   - `grafana-client-secret`: random 32-char hex (`openssl rand -hex 16`)
2. Build container: `mise run container-tag-and-release dex v1.0.0`

## Deployment sequence
1. Build container: `mise run container-tag-and-release dex v1.0.0`
2. Deploy Caddy: `mise run provision-indri -- --tags caddy`
3. Sync ArgoCD: `argocd app sync apps` → `argocd app sync dex`
4. Verify Dex: `curl https://dex.ops.eblu.me/.well-known/openid-configuration`
5. Sync Grafana: `argocd app sync grafana-config` → `argocd app sync grafana`
6. Test SSO: Visit `https://grafana.ops.eblu.me/login`, click "Sign in with Dex"

## Verification
- [ ] Container image exists: `mise run container-list` shows `dex:v1.0.0-nix`
- [ ] `curl https://dex.ops.eblu.me/.well-known/openid-configuration` returns valid OIDC discovery
- [ ] `curl https://dex.ops.eblu.me/healthz` returns healthy
- [ ] Grafana login shows "Sign in with Dex" button alongside local login
- [ ] OIDC flow: click Dex → enter credentials → redirect back → logged in as Admin
- [ ] Break-glass: local admin login still works
- [ ] `mise run services-check` passes

## Files changed
| File | Action | Purpose |
|------|--------|---------|
| `containers/dex/default.nix` | Create | NixOS container build |
| `argocd/apps/dex.yaml` | Create | ArgoCD app targeting ringtail |
| `argocd/manifests/dex/*` (8 files) | Create | K8s manifests (RBAC, ExternalSecret, Deployment, Service, Ingress) |
| `argocd/manifests/grafana-config/external-secret-dex-oauth.yaml` | Create | Grafana OIDC client secret |
| `argocd/manifests/grafana-config/kustomization.yaml` | Modify | Add new ExternalSecret resource |
| `argocd/manifests/grafana/values.yaml` | Modify | Add `auth.generic_oauth` config + envFromSecrets |
| `ansible/roles/caddy/defaults/main.yml` | Modify | Add `dex.ops.eblu.me` reverse proxy entry |
| `docs/changelog.d/feature-dex-oidc.feature.md` | Create | Changelog fragment |

Reviewed-on: https://forge.ops.eblu.me/eblume/blumeops/pulls/222
2026-02-19 20:24:24 -08:00

108 lines
2.7 KiB
YAML

# Grafana Helm values for blumeops
# Chart: https://github.com/grafana/helm-charts/tree/main/charts/grafana
# Admin credentials from pre-created secret
# Secret must exist before deploying - see grafana-config/README.md
admin:
existingSecret: grafana-admin
userKey: admin-user
passwordKey: admin-password
# Environment variables from secrets (for datasource credentials)
envFromSecrets:
- name: grafana-teslamate-datasource
optional: true
- name: grafana-dex-oauth
optional: true
# Persistence with PVC for SQLite database
persistence:
enabled: true
type: pvc
size: 1Gi
accessModes:
- ReadWriteOnce
# Grafana configuration via grafana.ini
grafana.ini:
server:
root_url: https://grafana.ops.eblu.me
security:
# Embedding disabled - iframe approach didn't work well for Homepage
allow_embedding: false
auth.generic_oauth:
enabled: true
name: Dex
client_id: grafana
client_secret: $__env{GF_AUTH_GENERIC_OAUTH_CLIENT_SECRET}
scopes: openid profile email
auth_url: https://dex.ops.eblu.me/auth
token_url: https://dex.ops.eblu.me/token
api_url: https://dex.ops.eblu.me/userinfo
allow_sign_up: true
role_attribute_path: "'Admin'"
auto_login: false
analytics:
check_for_updates: false
reporting_enabled: false
# Datasources - point to k8s-internal services
datasources:
datasources.yaml:
apiVersion: 1
datasources:
- name: Prometheus
type: prometheus
access: proxy
orgId: 1
uid: prometheus
url: http://prometheus.monitoring.svc.cluster.local:9090
isDefault: true
editable: false
- name: Loki
type: loki
access: proxy
orgId: 1
uid: loki
url: http://loki.monitoring.svc.cluster.local:3100
editable: false
- name: TeslaMate
type: postgres
access: proxy
orgId: 1
uid: TeslaMate
url: blumeops-pg-rw.databases.svc.cluster.local:5432
database: teslamate
user: teslamate
editable: false
jsonData:
sslmode: disable
maxOpenConns: 5
maxIdleConns: 2
connMaxLifetime: 14400
secureJsonData:
password: $TESLAMATE_DB_PASSWORD
# Dashboard provisioning - sidecar watches for ConfigMaps with label
sidecar:
dashboards:
enabled: true
label: grafana_dashboard
labelValue: "1"
folderAnnotation: grafana_folder
provider:
foldersFromFilesStructure: false
# Service configuration (Ingress will handle external access)
service:
type: ClusterIP
port: 80
# Resource limits for minikube
resources:
requests:
memory: "128Mi"
cpu: "100m"
limits:
memory: "512Mi"
cpu: "500m"