Environments

Manage multiple deployment environments with per-environment configuration and defaults.

Overview

Vibecarbon supports multiple named environments. Each environment is an independent deployment with its own infrastructure, database, secrets, and configuration.

vibecarbon deploy prod      # Production
vibecarbon deploy staging   # Staging
vibecarbon deploy dev       # Dev / preview

The environment name is free-form, so use any name that fits your workflow. prod (and its alias production) is the only special name and unlocks production-grade defaults.

Environment Defaults

The following defaults differ between prod and all other environments:

Setting	`prod`	`staging` / `dev` / `qa`
Backup schedule	Every 6 hours	Daily at 2 AM
Backup retention	30 days	7 days

The default server type does not vary by environment. It is derived from your provider's live catalog for the region you deploy into, filtered to the SKUs that region can actually place. Providers sell different generations in different locations (on Hetzner the EU locations and the US locations are on different shared-vCPU lines), so the default you get depends on your region.

If the CLI cannot reach the provider API it falls back to a built-in list, which is a best-effort snapshot rather than a guarantee; run the deploy prompt with a valid API token to select from live availability.

All settings can be overridden interactively during the deploy prompt. After the first deploy, choices are saved to .vibecarbon.json and reused automatically on subsequent deploys.

Managing Multiple Environments

You can run as many environments as you need. A typical setup:

Environment	Purpose
`prod`	Live traffic, production data
`staging`	Pre-release validation, mirrors prod config
`dev`	Feature development, throwaway data

Each environment gets its own cluster, domain, and database. They share nothing at the infrastructure level.

Node Scaling

Worker nodes are bounds-controlled at deploy time. The default is min=1, max=3 (cluster-autoscaler scales 0..(max-min) workers on top of the static floor). Worker bounds are set during the vibecarbon deploy interactive prompt or in .vibecarbon.json.

# Pick mode interactively, including k8s/k8s-ha and worker bounds
vibecarbon deploy prod

# Scripted: explicit mode, region, and skip prompts (hel1 is a Hetzner region)
vibecarbon deploy prod -mode k8s-ha -region hel1 -y

# A new environment on a different cloud (nyc3 is a DigitalOcean region)
vibecarbon deploy staging -provider digitalocean -region nyc3 -mode compose -y

-provider selects the cloud for a new environment, and -region takes that provider's own region ids. An environment keeps the provider it was first deployed with, so later deploys of the same environment don't repeat the flag. See Choosing a provider for which modes and regions each cloud offers.

Available modes: compose, compose-ha (multi-region with streaming replication), k8s, and k8s-ha. compose is free on Graphite; compose-ha, k8s, and k8s-ha require a Fullerene license. Mode availability varies by provider: k8s-ha runs on Hetzner and DigitalOcean today.

To re-tune bounds or worker server types on an existing cluster, run vibecarbon scale, and the wizard walks you through what's changeable. Power users can seed a server-type change with -type, using an id from that environment's provider catalog (e.g. vibecarbon scale prod -type cx33 -y on Hetzner).

Backups

Backups run automatically on a schedule that differs by environment:

# Create a manual backup
vibecarbon backup staging

# List backups
vibecarbon backup prod -l

# Restore (interactive picker)
vibecarbon restore staging

See the Deployment guide for full backup and restore options.

Destroying an Environment

To tear down an environment and all its infrastructure:

vibecarbon destroy staging

This removes the cluster, load balancers, volumes, and DNS records for that environment. Production environments prompt for confirmation before proceeding.