Garden vs KIWI NG

Garden is a DevOps automation tool for developing and testing Kubernetes applications. It creates production-like environments for development, testing, and CI on demand, then represents infrastructure, builds, services, and tests in a dependency-aware action graph. Graph-aware caching avoids rerunning unchanged images and tests. Garden Core uses YAML declarations to make workflows reproducible across developer machines and CI, and sync mode live-syncs changes to running deployments. Plugins cover Kubernetes, Helm, Terraform, Pulumi, and local scripts. Its Terraform provider validates and provisions infrastructure and exposes stack outputs to Garden actions; the Pulumi plugin is experimental and deploys stacks in dependency order. Garden Cloud's Remote Container Builder uses managed remote compute and shared Docker layer caches. Garden supports Linux, macOS, Windows, and self-hosted use. It requires Git 2.14 or newer, and local service builds need a local Kubernetes installation. The free tier has limits on build minutes, cache hits, and cache retention. Enterprise is a self-hosted Garden Cloud distribution installed into a customer's Kubernetes infrastructure.

KIWI NG is a free, daemonless command-line image builder for Linux. It turns versioned image descriptions and package repositories into bootable, reproducible artifacts. A prepared root tree can produce disk images, ISO files, or containers for virtual machines, cloud platforms, removable media, and specialized targets. Descriptions define metadata, repositories, packages, users, storage, target configuration, and image types, using config.xml or a .kiwi file. Overlays, archives, and scripts add customization. Documented package managers include apk, apt, dnf4, dnf5, microdnf, pacman, and zypper. Supported distributions include Fedora, CentOS Stream, openSUSE, Ubuntu, Debian, and Arch Linux. Builds support multiple architectures and a local cache backend. Installation is available through distribution packages, Open Build Service, or pip. A recent Linux host, Python 3.9 or newer, Git, and at least 15 GB recommended free disk space are required. QEMU can optionally boot and inspect outputs; GPL-3.0 governs the source.