Kubernetes Deployment

We use a custom Helm chart to deploy the staging environment on the NIRD Kubernetes cluster.

Architecture

The deployment splits the monolithic stack into cloud-native components:

1. Specify Backend (backend): Runs the Django application (Gunicorn).
   • Uses an InitContainer to copy static assets (.js, .css) from the official image to a Shared PVC.
   • Mounts configuration from a Kubernetes Secret.
2. Static Server (static): A lightweight Nginx service.
   • Mounts the Shared PVC to serve the assets extracted by the backend.
   • Handles path rewrites (e.g., /static/config/ -> generated settings).
3. Worker (worker): Runs the Celery worker for asynchronous tasks.
4. Ingress: Routes traffic:
   • /static/* -> Static Server
   • /* -> Specify Backend
5. MariaDB: A dedicated MariaDB instance running as a sub-chart dependency (Bitnami).

Prerequisites

• Kubernetes Cluster (v1.19+)
• ReadWriteMany Storage Class (Required for shared assets)
• Ingress Controller
• Helm 3.x
• Bitnami Repo: helm repo add bitnami https://charts.bitnami.com/bitnami

Quick Start

1. Build Dependencies

The chart depends on the bitnami/mariadb chart. You must build this dependency first.

helm dependency build charts/specify7

2. Configure Credentials

Edit charts/specify7/values.yaml to set your passwords and secrets.

3. Deploy

helm install staging ./charts/specify7 \
  --set ingress.enabled=true \
  --set ingress.hosts[0].host=specify.yourdomain.com

Configuration

Setting	Description	Default
persistence.storageClass	Critical: Must support RWX	"" (defaultSC)
mariadb.enabled	Enable the bundled MariaDB	true
specify.database	Database connection (auto-configured if mariadb enabled)	mariadb
specify.secretKey	Django Secret Key	change-me...