mirror of
https://github.com/stablyai/orca.git
synced 2026-09-30 00:03:15 +00:00
Copies the private repository's relay side: the IAM-only fence broker, the operations console and incident monitor, the relay Terraform root with its backend configuration and tfvars, and the deploy/capacity/admission/rehome/ monitoring scripts the workflows call, with their contract tests, contracts, and fixtures. The foundation and apps Terraform roots and the API and auth services stay private. Four surfaces that spanned both trees are narrowed to the relay side rather than left with a dangling read: the infra runner and the root-partition and workload-identity-condition renderers now declare only the relay root, and the Cloud SQL rollout census drops the six app workflows that are not here.
16 lines
1.2 KiB
JSON
16 lines
1.2 KiB
JSON
{
|
|
"comment": "Production Cloud SQL consumers owned by the orca-cloud application tree. The relay tree ships without apps/auth and apps/api, so the values it needs for the connection budget are published here instead of parsed out of application source. dev/scripts/production-cloud-sql-app-consumers.test.mjs binds every field back to its source with the regexes the budget script used, so application-side drift fails application-side CI.",
|
|
"authInstances": 2,
|
|
"authPoolMax": 10,
|
|
"apiInstances": 10,
|
|
"apiPoolMax": 5,
|
|
"maxConnections": 400,
|
|
"sources": {
|
|
"authInstances": "infra/terraform-apps/environments/production.tfvars: auth_max_instances",
|
|
"authPoolMax": "apps/auth/src/db/database.ts: new pg.Pool({ ..., max: N })",
|
|
"apiInstances": "infra/terraform-apps/environments/production.tfvars: max_instances (the artifact API service; the first max_instances not prefixed auth_)",
|
|
"apiPoolMax": "apps/api/src/skills/skill-database.ts: new pg.Pool({ ..., max: N })",
|
|
"maxConnections": "Tier default for google_sql_database_instance.auth; infra/terraform-foundation/database.tf declares no max_connections database flag that would override it"
|
|
}
|
|
}
|