O.0 prep from OPERATOR-PLAN.md. Mechanical refactor before adding partner
management and operator-specific endpoints. The service now owns more than
just provisioning orchestration (it'll soon own partners, tenant lifecycle
actions, multi-audience JWT validation), so the name 'platform-api' reflects
its scope better.
What changed:
- Directory: services/provisioning/ -> services/platform-api/
- Package: @dezky/provisioning -> @dezky/platform-api
- Docker: container_name dezky-provisioning -> dezky-platform-api;
compose service key 'provisioning' -> 'platform-api'; volume
provisioning_node_modules -> platform_api_node_modules
- Portal: PROVISIONING_INTERNAL_URL env var -> PLATFORM_API_INTERNAL_URL,
default URL http://provisioning:3001 -> http://platform-api:3001 in all
three proxy routes (me.get.ts, tenants/index.post.ts, tenants/[slug]/
reconcile.post.ts), plus NUXT_API_BASE updated
- Health endpoint service identifier and main.ts log lines updated to
'dezky-platform-api'
- Docs swept: README, CLAUDE.md, SERVICES.md, AUTHENTIK-SETUP.md,
NEXT-STEPS.md, TROUBLESHOOTING.md, OPERATOR-PLAN.md, traefik/dynamic.yml
What deliberately stays:
- Internal module names ProvisioningService / ProvisioningModule (those
describe an orchestration sub-concern, not the service's purpose)
- Tenant.provisioningStatus / provisioningErrors field names (state
per integration, not service name)
- File services/platform-api/src/tenants/provisioning.service.ts
- 'Hetzner provisioning' references in production-prep docs (infrastructure
provisioning, unrelated)
Verified end-to-end after rename: /api/me returns 200 with profile + 2
tenants + subscription, /api/tenants/dezky/reconcile returns 200 with
Authentik integration still ok.
OPERATOR-PLAN.md O.0 checkboxes ticked.
Phase 4 from docs/NEXT-STEPS.md. POST /tenants now writes Mongo AND drives
external service provisioning. A new POST /tenants/:slug/reconcile endpoint
retries the orchestration — useful when an upstream was down at create time
or external state drifted out of band.
Integration clients (services/provisioning/src/integrations/):
- AuthentikClient: real implementation. ensureGroup() is idempotent — looks
up the group by name, creates if missing, returns either way. Group
attributes record the tenant slug + Mongo id so we can trace back
- StalwartClient: stubbed. v0.16 removed the REST management API in favor
of JMAP, which is significantly more work to wrap. TODO comment points
to https://stalw.art/docs/api/management/overview for the follow-up
- OcisClient: stubbed. Needs libregraph /drives endpoint with service-to-
service auth via OIDC client_credentials
Orchestration (provisioning.service.ts):
- Each step runs independently; one failure doesn't roll back the others
- Per-step state recorded on Tenant.provisioningStatus (ok/skipped/error/
pending) plus error message on Tenant.provisioningErrors
- Steps return their own terminal state — 'skipped' for stubs, void
defaults to 'ok' for real integrations
- Mongoose markModified() required for nested subdoc mutations to persist
- Tenant auto-flips status: pending → active when all steps are ok|skipped
Portal proxy routes (apps/portal/server/api/tenants/):
- POST /api/tenants and POST /api/tenants/:slug/reconcile forward the
signed-in user's access token to the provisioning service. Lets the
browser drive provisioning without minting tokens by hand. Will be
replaced by a real "create workspace" flow with UI later
docker-compose: AUTHENTIK_API_URL/STALWART_API_URL/OCIS_API_URL now point
at the public Traefik-routed hostnames (with mkcert CA mounted into the
provisioning container so Node fetch trusts them). Previously these
pointed at internal Docker hostnames which doesn't work for Authentik
because of TLS issuer mismatch against the JWT.