Skip to main content

Spoolman on GKE Autopilot

Spoolman on GKE Autopilot

Spoolman is a free, open-source inventory and usage tracker for 3D-printing filament spools — vendors, materials, remaining weight, cost per spool, and per-print consumption. It ships as a single-process Python/FastAPI backend with a bundled static Vue/Quasar frontend. This module deploys Spoolman on GKE Autopilot on top of the App_GKE foundation, which provisions and manages the shared Google Cloud and Kubernetes infrastructure.

This guide focuses on the cloud services Spoolman uses and how to explore and operate them from the Google Cloud Console and the command line. For the mechanics common to every GKE application — Workload Identity, ingress, autoscaling, CI/CD, Cloud Armor, IAP, Binary Authorization, VPC Service Controls, backups, and the deployment lifecycle — refer to the App_GKE foundation guide rather than repeating them here.


1. Overview

Spoolman runs as a single Python/FastAPI pod — there is no separate frontend workload; the Vue/Quasar UI is bundled and served from the same process. The deployment wires together a minimal set of Google Cloud services:

CapabilityGoogle Cloud serviceNotes
ComputeGKE AutopilotPrebuilt ghcr.io/donkie/spoolman image, 1 vCPU / 512Mi by default
DatabaseCloud SQL for PostgreSQL 15Required — this module standardises on Postgres (Spoolman upstream also supports MySQL/SQLite/CockroachDB)
Object storageNoneSpoolman keeps all state in Postgres; no GCS bucket is provisioned
CacheNoneSpoolman has no Redis/cache integration
SecretsSecret ManagerOnly the auto-generated database password — Spoolman has no admin/API-key bootstrap secret of its own
IngressCloud Load BalancingExternal LoadBalancer by default

Sensible defaults worth knowing up front:

  • PostgreSQL 15 is the only supported engine in this module. database_type is fixed by Spoolman_Common; Spoolman upstream also supports MySQL and CockroachDB via env vars, but this module does not expose that choice.
  • No custom build. container_image_source = "prebuilt" deploys ghcr.io/donkie/spoolman directly — there is no Dockerfile, no Cloud Build step.
  • No init job. The Foundation auto-creates the Postgres role and database; Spoolman runs its own Alembic migrations automatically on every container start.
  • No application secrets. Spoolman ships with no authentication at all — whoever can reach the Service has full read/write access to the inventory. There is no login gate to bootstrap and nothing generated in Secret Manager beyond the database password. If that is not acceptable, put the service behind IAP (enable_iap = true) or a Cloud Armor IP allowlist.
  • service_type = "LoadBalancer" by default. Spoolman is a browser-driven web UI, so the Service is externally reachable out of the box.
  • reserve_static_ip = false by default. Spoolman bakes no self-referencing URL into boot-time config (only SPOOLMAN_CORS_ORIGIN matters, and only for cross-domain access), so this module conserves the project's often-tight static-IP quota by not reserving one.
  • Connections use the cloud-sql-proxy sidecar's loopback, not TCP. Spoolman's SQLAlchemy layer builds its connection via URL.create() (a structured object, not string concatenation), so 127.0.0.1 passes through cleanly with no URL-parsing issue and no TLS/sslmode configuration needed.

2. Google Cloud Services & How to Explore Them

All commands assume you have run gcloud container clusters get-credentials <cluster> --region <region> --project <project> and that PROJECT, REGION, and NAMESPACE are set. The namespace and other identifiers are reported in the deployment Outputs.

A. GKE Autopilot — the Spoolman workload

Spoolman runs as a single Deployment/pod on Autopilot, which bills for the CPU/memory the pod actually requests.

  • Console: Kubernetes Engine → Workloads → select the Spoolman workload for pods, revisions, and events. Kubernetes Engine → Services & Ingress shows the external IP.
  • CLI:
    kubectl get pods,svc -n "$NAMESPACE"
    kubectl logs -n "$NAMESPACE" deploy/<service-name> --tail=100

See App_GKE for how Autopilot and scaling are managed.

B. Cloud SQL for PostgreSQL 15

Spoolman stores all inventory data (spools, filaments, vendors, usage history) in a managed Cloud SQL for PostgreSQL 15 instance. The pod reaches it privately through the Cloud SQL Auth Proxy sidecar over loopback; no public IP is exposed. There is no initialization job — Spoolman applies its own schema migrations on every boot.

  • Console: SQL → select the instance for connections, backups, flags, and metrics.
  • CLI:
    gcloud sql instances list --project "$PROJECT"
    gcloud sql instances describe <instance-name> --project "$PROJECT"
    gcloud sql connect <instance-name> --user=<db-user> --database=<db-name> --project "$PROJECT"

The instance name, database, user, and password secret are in the Outputs. See App_GKE for the connection model, backups, and password rotation.

C. Secret Manager

Only the auto-generated database password lives in Secret Manager — Spoolman has no admin account or API key of its own to bootstrap.

  • Console: Security → Secret Manager.
  • CLI:
    gcloud secrets list --project "$PROJECT" --filter="name~spoolman"
    gcloud secrets versions access latest --secret=<db-password-secret> --project "$PROJECT"

D. Networking & ingress

The Service is reachable at its external LoadBalancer IP by default (service_type = "LoadBalancer", reserve_static_ip = false so the IP is ephemeral unless changed).

  • Console: Kubernetes Engine → Services & Ingress.
  • CLI:
    kubectl get svc -n "$NAMESPACE" -o wide

See App_GKE for custom domains, CDN, and Cloud Armor.

E. Cloud Logging & Monitoring

Container logs flow to Cloud Logging; GKE and Cloud SQL metrics flow to Cloud Monitoring, with optional uptime checks and alert policies.

  • Console: Logging → Logs Explorer; Monitoring → Dashboards / Alerting.
  • CLI:
    kubectl logs -n "$NAMESPACE" deploy/<service-name> --tail=100

3. Spoolman Application Behaviour

  • No first-deploy database setup job. Unlike most application modules in this catalogue, Spoolman needs no db-init job — the Foundation creates the Postgres role and database, and Spoolman's own Alembic migrations run automatically on every container start (including the very first boot).
  • No authentication. There is no login page, no admin account, and no API key gate. Anyone who can reach the Service can view and modify the entire inventory. Decide your access-control approach (IAP, Cloud Armor allowlist, or accept public read/write) before exposing the LoadBalancer IP.
  • Health path. /api/health is public and unauthenticated, returning a 200/OK JSON status once the server (and its DB connection) is up. Both the startup and liveness probes target this path.
  • Database engine locked to Postgres. Spoolman's own SPOOLMAN_DB_TYPE environment variable selects the engine; this module always sets it to postgres. Never unset it via environment_variables — without it, Spoolman silently falls back to a throwaway container-local SQLite file with no error at all.
  • Inspect Cloud SQL connectivity:
    kubectl exec -n "$NAMESPACE" deploy/<service-name> -- env | grep -i spoolman_db

4. Configuration Variables

Variables are grouped exactly as they appear on the deployment platform. Only settings specific to or notable for Spoolman are listed; every other input is inherited from App_GKE with its standard behaviour.

Group 3 — Application Identity

VariableDefaultDescription
application_namespoolmanBase name for resources. Do not change after first deploy.
application_display_nameSpoolmanHuman-readable name.
application_versionlatestImage tag pulled from ghcr.io/donkie/spoolman. Genuinely prebuilt — no Dockerfile/build-arg pinning concerns.

Group 4 — Container & Scale

VariableDefaultDescription
deploy_applicationtrueSet false to provision infrastructure only.
container_image_sourceprebuiltForwarded to the Foundation — required, or the default "custom" silently triggers a Kaniko build attempt with no Dockerfile.
container_port8000Spoolman's default listen port.
container_resources{ cpu_limit = "1000m", memory_limit = "512Mi" }Ample for a single-tenant filament tracker.
min_instance_count / max_instance_count0 / 1Scale-to-zero is safe — Spoolman has no background work.

Group 16 — Database

VariableDefaultDescription
db_host_env_var_nameSPOOLMAN_DB_HOSTAliases the Foundation's DB_HOST (127.0.0.1 via the cloud-sql-proxy sidecar on GKE).
db_user_env_var_nameSPOOLMAN_DB_USERNAMEAliases DB_USER.
db_password_env_var_nameSPOOLMAN_DB_PASSWORDAliases DB_PASSWORD.
db_name_env_var_nameSPOOLMAN_DB_NAMEAliases DB_NAME.
db_port_env_var_nameSPOOLMAN_DB_PORTAliases DB_PORT.
application_database_name / application_database_userspoolman / spoolmanImmutable after first deploy.

Group 14 — Storage & Filesystem

VariableDefaultDescription
create_cloud_storagefalseNo GCS bucket needed — all state lives in Cloud SQL.
enable_nfsfalseNo shared filesystem needed.

Group 19 — Networking

VariableDefaultDescription
service_typeLoadBalancerSpoolman is a browser-driven web UI, exposed externally by default.
reserve_static_ipfalseConserves the project's static-IP quota; Spoolman bakes no self-referencing URL into boot-time config.

Group 10 — Probes & Lifecycle

VariableDefaultDescription
startup_probe_configHTTP /api/health, 10s delayPublic, unauthenticated.
health_check_configHTTP /api/health, 30s periodPublic, unauthenticated.

All other inputs (CI/CD, backups, VPC-SC, Cloud Armor, IAP, Redis, stateful PVCs) are inherited from App_GKE with standard behaviour — Spoolman uses none of them by default.


5. Outputs

OutputDescription
service_nameKubernetes Service name.
namespaceKubernetes namespace.
service_cluster_ipInternal ClusterIP.
service_external_ipExternal LoadBalancer IP.
service_urlFull URL of the deployed workload.
database_instance_nameCloud SQL instance name.
database_name / database_userApplication database name / user.
database_password_secretSecret Manager secret holding the DB password.
database_host / database_portDB endpoint / port.
storage_bucketsAlways empty — Spoolman needs no GCS bucket.
container_image / container_registryDeployed image and Artifact Registry repo (when mirroring is enabled).
monitoring_enabledMonitoring status.
initialization_jobsAlways empty — Spoolman needs no init job.
kubernetes_readyWhether the workload reports ready.
deployment_id / tenant_id / resource_prefixNaming identifiers.
project_id / project_numberProject identifiers.
vpc_sc_enabled / vpc_sc_perimeter_name / vpc_sc_dry_run_modeVPC-SC status.
audit_logging_enabled / artifact_registry_cmek_enabledAudit logging and CMEK status.

6. Configuration Pitfalls & Sensible Defaults

Risk: Critical (data loss / outage / security) — High (service degraded) — Medium (cost or partial degradation) — Low (minor).

SettingSensible valueRiskConsequence if wrong
No authentication (built-in)Front with IAP or Cloud Armor if neededCriticalAnyone who can reach the Service can read and modify the entire filament inventory — there is no login gate to disable.
SPOOLMAN_DB_TYPE (auto-injected postgres)Never unset via environment_variablesCriticalUnsetting it silently falls back to a throwaway container-local SQLite file — no error, and all data is lost on every pod restart.
application_database_name / application_database_userSet onceCriticalImmutable after first deploy; renaming recreates the DB/user and destroys all data.
container_image_sourceprebuilt (do not override to custom)CriticalSetting "custom" triggers a Kaniko build attempt against a module with no Dockerfile — the build fails outright.
service_typeLoadBalancer (default)HighSwitching to ClusterIP makes the service unreachable from a browser without kubectl port-forward or a separate ingress.
SPOOLMAN_DB_QUERYLeave empty unless troubleshootingMediumEscape hatch for a TCP + sslmode fallback — only needed if the loopback connection path is ever found unreliable; not required for normal operation.
reserve_static_ipfalse (default)LowSet true only if you need a stable IP for DNS/firewall allowlisting — the project's static-IP quota is limited and shared across the tenant.
min_instance_count0 (default)LowSpoolman has no background work, so scale-to-zero is safe.

For the foundation behaviour referenced throughout — Workload Identity, scaling, ingress and load balancing, CI/CD, Cloud Armor, IAP, Binary Authorization, VPC-SC, backups, and image mirroring — see App_GKE. Spoolman-specific application configuration shared with the Cloud Run variant is described in Spoolman_Common.