4.1.9. Bootstrap Scripts

Cluster provisioning, Kubernetes secrets injection, and database initialization scripts

The bootstrap/ directory contains three Python scripts that prepare a fresh environment for Armada. They are meant to be executed in order: secrets first, then cluster resources, then database.

Prerequisites

Install the Python dependencies before running any script:

pip install -r bootstrap/requirements.txt
Package Version Used by
kubernetes 35.0.0 bootstrap_secrets.py
pymssql 2.3.13 bootstrap_database.py

All scripts read configuration from the .env file at the project root.

1. bootstrap_secrets.py

Creates or updates Kubernetes secrets in a target namespace from values defined in .env.

Secrets managed

Secret name Type Keys consumed
armada-docker-username-secret Opaque DOCKER_HUB_USERNAME
armada-docker-registry-secret kubernetes.io/dockerconfigjson DOCKER_HUB_USERNAME, DOCKER_HUB_PASSWORD, DOCKER_HUB_MAIL
armada-ipqs-secret Opaque IPQS_KEY
armada-sql-server-secret Opaque SQL_SERVER_USER, SQL_SERVER_PASSWORD, SQL_SERVER_DB, SQL_SERVER_NAME

If any required key is missing from .env, the corresponding secret is skipped (not treated as an error). The script prints a summary of how many secrets were applied versus skipped.

Idempotent behaviour

The script checks whether each secret already exists in the namespace. Existing secrets are replaced; missing ones are created.

Usage

Normally :

python bootstrap/bootstrap_secrets.py

or if needed

python bootstrap/bootstrap_secrets.py --namespace default --env-file .env
Flag Default Description
--namespace default Kubernetes namespace to write secrets into
--env-file — Path to the .env file

2. bootstrap_cluster_resources.py

Builds Docker images (when applicable), then installs the Armada Helm chart into the active Kubernetes cluster. The secrets created in step 1 must already exist, since the Helm chart Deployments reference them via secretKeyRef and imagePullSecrets.

Installation modes

The script presents an interactive menu with three options:

Mode When to use What it does
Public Production / CI with official images Runs helm install pointing to the pre-built armadasvc Docker Hub images. No local build step.
Private Production with a custom registry Builds all six service images, pushes them to your Docker Hub account, then runs helm install with imagePullSecrets.
Minikube Local development Builds images inside Minikube’s Docker daemon (eval $(minikube docker-env)), sets imagePullPolicy=Never, then runs helm install.

Service images

Private and Minikube modes build one image per service:

Source directory Image name
services/agent armada-agent
services/backend armada-backend
services/fingerprint-provider armada-fingerprint-provider
services/frontend armada-frontend
services/orchestrator armada-orchestrator
services/proxy-provider armada-proxy-provider

Required variables

It fetches from .env :

  • DOCKER_HUB_USERNAME — required for Private and Minikube modes.
  • DOCKER_HUB_PASSWORD — required for Private mode (registry authentication).

Example

python bootstrap/bootstrap_cluster_resources.py
# Follow the interactive prompt to choose a mode.

After completion the script prints a summary with useful verification commands (kubectl get pods, helm list, etc.).

3. bootstrap_database.py

Connects to SQL Server and creates the five tables Armada requires. Every CREATE TABLE statement is guarded with IF NOT EXISTS, making the script safe to run multiple times.

Tables

For the complete database schema (tables, columns, types, relationships), see the Database Reference

Usage

python bootstrap/bootstrap_database.py

1. bootstrap_secrets.py             Inject secrets into the cluster
              │
              ▼
2. bootstrap_cluster_resources.py   Build images & helm install
              │
              ▼
3. bootstrap_database.py            Create SQL Server tables

Run the scripts in this order to go from an empty cluster to a fully provisioned Armada environment.