> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aperium.apps.hillspire.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Prerequisites

> What must exist before applying the GCP deployment.

The GCP deployment starts at the shared environment bootstrap layer. The items below must already be in place before you apply it.

## GCP and identity

* A GCP project for the shared apps environment.
* A Terraform Cloud / HCP Terraform organization and workspaces.
* Credentials that can run Terraform against GCP.
* Authority to delegate the parent DNS zone to the managed subdomain returned by Terraform.

## Source control and ArgoCD

* A Git repository that holds your extracted deployment layout.
* A GitHub App that ArgoCD can use for repository access. You will need:
  * `github_app_id`
  * `github_app_installation_id`
  * `github_app_private_key` (loaded as a sensitive Terraform variable in the shared env workspace)

## First values to replace

Before applying the deployment as a live repo, change these values:

<Steps>
  <Step title="Repo URL placeholder">
    Replace the placeholder `https://github.com/YOUR_ORG/YOUR_REPO.git` everywhere it appears. ArgoCD app manifests and bootstrap Terraform files already point at this placeholder.
  </Step>

  <Step title="Terraform variables">
    Start from `envs/aperium-apps-prod/tf/vars.auto.tfvars.example` and `apps/aperium/envs/prod/tf/vars.auto.tfvars.example`. Compare with `vars.reference.tfvars` in each directory for the extracted reference values.
  </Step>

  <Step title="GCP, TFC, and DNS placeholders">
    Throughout values files, replace at minimum:

    * `YOUR_GCP_PROJECT_ID`
    * `YOUR_GCP_REGION`
    * `YOUR_GCP_ZONE`
    * `YOUR_DOMAIN`
    * `YOUR_CLUSTER_SECRET_STORE_NAME`
    * `YOUR_TFC_ORG`
    * `YOUR_SHARED_ENV_WORKSPACE`
    * `YOUR_APP_WORKSPACE`
    * `YOUR_PREFECT_CLOUDSQL_INSTANCE`
  </Step>

  <Step title="Secret payloads">
    Load the payloads described in [Secrets](/deployment/gcp/secrets). Terraform creates the Secret Manager *containers*; in most cases you still need to populate the values yourself.
  </Step>
</Steps>

## Prefect prerequisite checklist

Before syncing the `prefect` application, verify that all of the following are true:

* `YOUR_PREFECT_CLOUDSQL_INSTANCE` is set and reachable from the cluster.
* The Prefect runtime GSA exists, for example `prefect@YOUR_GCP_PROJECT_ID.iam.gserviceaccount.com`.
* `prefect-admin-credentials` exists in the external secret store and produces a Kubernetes secret that contains `auth-string`.
* `prefect-server.yaml` and `prefect-worker-aperium.yaml` values have been templated for your environment.
* You have a bootstrap path to create the `aperium-pool` work pool after the Prefect server is healthy.

## Conventions to remember

* `vars.reference.tfvars` files are reference snapshots. They are not auto-loaded.
* `vars.auto.tfvars.example` files are the starting point for real usage.
* The local ArgoCD app manifests are wired to your extracted deployment repo, not back to any upstream source repo.
* Many live prod-specific values have been replaced with placeholders, but the deployment structure is preserved.

## Quick validation

Before pushing the templated repo:

```bash theme={null}
terraform fmt -check -recursive .
rg -n "YOUR_ORG/YOUR_REPO|vars.reference.tfvars|YOUR_GCP_PROJECT_ID|YOUR_DOMAIN" .
find envs/aperium-apps-prod/argo -maxdepth 1 -type f | sort
```

The first command catches Terraform formatting drift. The second surfaces any placeholders you forgot to replace. The third confirms the ArgoCD app set is intact.