Clanker Docs

Appearance

Google Cloud Platform (GCP) ​

Clanker supports querying GCP infrastructure using the gcloud CLI. Queries are routed to GCP either by an explicit flag or by automatic keyword detection.

Authentication ​

GCP authentication uses the gcloud CLI and its standard credential chain.

Setting up authentication ​

bash
# Login to GCP
gcloud auth login

# Set the default project
gcloud config set project your-project-id

# For the Gemini (Vertex AI) provider, also run:
gcloud auth application-default login

Configuration ​

Add your GCP project ID to ~/.clanker.yaml:

yaml
infra:
  gcp:
    project_id: your-gcp-project-id

Project ID resolution order ​

  1. --gcp-project CLI flag
  2. infra.gcp.project_id in the config file
  3. GCP_PROJECT environment variable
  4. GOOGLE_CLOUD_PROJECT environment variable
  5. GCLOUD_PROJECT environment variable

Querying GCP ​

Explicit flag ​

bash
clanker ask --gcp "List all Cloud Run services"

Automatic routing ​

If your question mentions GCP-specific services, Clanker routes to GCP automatically:

bash
clanker ask "What Cloud Run services are deployed?"

Combining with other providers ​

GCP context can be combined with other providers in a single query:

bash
clanker ask --gcp --terraform "Compare Terraform state with actual GCP resources"

Supported Services ​

Clanker queries GCP services through the gcloud CLI. Commonly queried services include:

  • Compute Engine (instances, instance groups)
  • Cloud Run (services, jobs, revisions, worker pools, domain mappings, multi-region services)
  • Google Kubernetes Engine (GKE clusters)
  • Cloud Storage (buckets)
  • Cloud SQL, AlloyDB, BigQuery, Spanner, Bigtable, Memorystore Redis and Memcached
  • Cloud Functions Gen 1 and Gen 2
  • Pub/Sub (topics, subscriptions)
  • VPC networking, load balancers, Cloud Armor, Cloud DNS, API Gateway
  • Workflows, Batch, Vertex AI endpoints and indexes, Firestore, Firebase apps
  • Artifact Registry repositories, packages, and Docker images
  • Cloud Build, Cloud Deploy, Eventarc, Scheduler, Tasks, Secret Manager, Cloud KMS, logging sinks, and alert policies

Direct inventory examples:

bash
clanker gcp list run-worker-pools --project my-project --location us-central1
clanker gcp list vertex-endpoints --project my-project --location us-central1
clanker gcp list eventarc-triggers --project my-project
clanker gcp list api-gateway --project my-project

Maker Pipeline ​

The maker feature supports GCP infrastructure provisioning:

bash
# Generate a plan
clanker ask --maker --gcp "Create a Cloud Run service called my-api from gcr.io/my-project/my-api:latest"

# Apply the plan
clanker ask --apply --plan-file plan.json

GCP maker plans are output directly without the enrichment step used for AWS plans.

Backend Credentials ​

For team environments, GCP credentials (including service account JSON) can be stored and retrieved through the Clanker backend:

bash
clanker credentials store gcp
clanker ask --api-key ck_live_abc123 --gcp "List Cloud Run services"