Skip to main content
Griffin Hub is the backend service that stores monitors, schedules runs, processes results, and serves the API. It’s what the CLI communicates with when you run griffin apply, griffin runs, or griffin metrics.

​
What the hub does

  • Stores monitors — Receives and persists monitor definitions from the CLI
  • Schedules execution — Triggers monitors based on their configured frequency
  • Processes results — Stores run outcomes and aggregates metrics
  • Manages secrets — Resolves secret references at execution time
  • Sends notifications — Evaluates notification rules and dispatches alerts
  • Serves the API — Provides REST endpoints for the CLI and integrations

​
Deployment modes

​
Standalone

The hub runs with a built-in executor in a single process. This is the simplest setup: 
cd griffin-hub
npm run start:standalone
One process handles everything: API, scheduling, and test execution.

​
Distributed

The hub runs as a control plane only, with separate executor agents: 
# Start the hub (control plane only)
cd griffin-hub
npm start

# Start agents separately (can be on different machines)
# Agents register with the hub and pull jobs from the queue
This mode supports multiple executors in different locations for geographic distribution and horizontal scaling.

​
Requirements

  • Node.js 18+
  • PostgreSQL 14+ (for storage and job queue)
  • Appropriate environment variables (see Configuration)

​
Quick start

# Set up the database
export DATABASE_URL="postgresql://user:pass@localhost:5432/griffin"
cd griffin-hub
npm run db:push

# Start in standalone mode
npm run start:standalone

Configuration

All environment variables and options.

Self-Hosting Guide

Step-by-step deployment instructions.