GCP deployment

For Google Cloud enterprise deployments, storyshelf server init offers a gcp deploy target that pins the full reference stack — one Cloud Run service (it serves the UI and runs the captures from the database queue, with CPU always allocated), GCS bucket, Cloud SQL Postgres (db-f1-micro shared-core default), Secret Manager, Identity Platform tenant, and optional Cloud DNS with automatic managed TLS via the domain mapping — and writes it as Terraform under terraform/.

Scaffold

storyshelf server init
# ? Deploy target? Google Cloud (Cloud Run + GCS + Postgres + Identity Platform)
# ? GCP project ID? my-gcp-project
# ? GCP region? us-central1
# ? Public domain for the app? (empty skips DNS)
# ? Identity Platform tenant display name? (empty skips it)

Apply

cd my-server
gcloud auth login
gcloud config set project my-gcp-project
terraform -chdir=terraform init
terraform -chdir=terraform plan -out=tfplan   # review before applying
terraform -chdir=terraform apply tfplan       # or: npm run infra:apply

Wiring outputs

Wire the outputs into env (terraform output -json / npm run infra:outputs): GCS_BUCKET + GOOGLE_CLOUD_PROJECT, DATABASE_URL (?sslmode=require), OIDC_*, SECRET. Output keys (run_url, gcs_bucket, database_url, identity_tenant_id) are a stable contract parsed by CI — always present. Grant the runtime service account the Cloud SQL Client and Secret Manager accessor roles post-apply — Terraform creates the resources, IAM bindings stay with your org policy.

See the generated terraform/README.md for the CI test profile (unique project prefix per run, documented destroy).

Pair with @storyshelf/storage-gcs and @storyshelf/db-postgres. The database is also the capture queue, so there is no topic, subscription or worker service to provision.

Related

• Deployment overview — Docker Compose, subdomains, auth, other clouds