keyhan
1572b3ce66
Build and Deploy Platform / build-and-deploy (push) Failing after 50m25s
Backend now fails production validation without CLUSTER_KUBECONFIG_KEY. Add cluster-kubeconfig-key to the chart secret and env, plus secrets.existingSecret so GitOps deployments can use a pre-created (sealed) Secret instead of the lookup/randAlphaNum template that churns under Argo CD's helm template rendering. Co-authored-by: Cursor <cursoragent@cursor.com>
cloudhost-platform
Helm chart for the CloudHost control plane: backend API, frontend UI, PostgreSQL, and Redis.
Quick install
helm upgrade --install cloudhost . \
-n cloudhost --create-namespace \
-f values-production.example.yaml
Ingress / TLS
| Value | Description |
|---|---|
ingress.enabled |
Create Ingress resources |
ingress.tls.enabled |
Enable cert-manager TLS |
ingress.tls.clusterIssuer |
ClusterIssuer name (e.g. letsencrypt-prod) |
ingress.frontend.host |
UI hostname |
ingress.api.host |
API hostname |
Requires NGINX Ingress Controller and cert-manager when TLS is enabled.
Frontend image
Build with the public API URL baked in:
docker build -t $REG/cloudhost-frontend:tag \
--build-arg NEXT_PUBLIC_API_URL=https://api.platform.example.com \
../../frontend
Migrations
SQL files in migrations/ run via a post-install/post-upgrade Job when migrations.enabled is true. They assume incremental schema changes on top of an existing database; for a completely empty database you may need to bootstrap schema first (e.g. one-time NODE_ENV=development or manual setup).