Genkit Infra Expert

Prerequisites

Before using this skill, ensure:

• Google Cloud project with Firebase enabled
• Terraform 1.0+ installed
• gcloud and firebase CLI authenticated
• Genkit application built and containerized
• API keys for Gemini or other AI models
• Understanding of Genkit flows and deployment options

Instructions

1. Choose Deployment Target: Firebase Functions, Cloud Run, or GKE
2. Configure Terraform Backend: Set up remote state in GCS
3. Define Variables: Project ID, region, Genkit app configuration
4. Provision Compute: Deploy functions or containers
5. Configure Secrets: Store API keys in Secret Manager
6. Set Up Monitoring: Create dashboards for token usage and latency
7. Enable Auto-scaling: Configure min/max instances
8. Validate Deployment: Test Genkit flows via HTTP endpoints

Output

Firebase Functions:

# {baseDir}/terraform/functions.tf
resource "google_cloudfunctions2_function" "genkit_function" {
  name     = "genkit-ai-flow"
  location = var.region

  build_config {
    runtime     = "nodejs20"
    entry_point = "genkitFlow"
  }

  service_config {
    max_instance_count = 100
    available_memory   = "512Mi"
    timeout_seconds    = 300
  }
}

Cloud Run Service:

resource "google_cloud_run_v2_service" "genkit_service" {
  name     = "genkit-api"
  location = var.region

  template {
    scaling {
      min_instance_count = 1
      max_instance_count = 10
    }
    containers {
      image = "gcr.io/${var.project_id}/genkit-app:latest"
      resources {
        limits = {
          cpu = "2"
          memory = "1Gi"
        }
      }
    }
  }
}

Error Handling

Build Failures

• Error: "Cloud Function build failed"
• Solution: Check package.json dependencies and Node.js runtime version

Cold Start Latency

• Warning: "High latency on first request"
• Solution: Set min_instance_count >= 1 to keep warm instances

Secret Access Denied

• Error: "Permission denied accessing secret"
• Solution: Grant secretAccessor role to Cloud Run/Functions service account

Memory Exceeded

• Error: "Container killed: out of memory"
• Solution: Increase available_memory or optimize Genkit flow memory usage

Resources

• Genkit Deployment: https://genkit.dev/docs/deployment
• Firebase Terraform: https://registry.terraform.io/providers/hashicorp/google/latest
• Genkit examples in {baseDir}/genkit-examples/