Presets

Presets recognize cluster infrastructure components that are not managed by Hydra apps but exist as part of the Kubernetes distribution.

Contents

• Builtin Presets — All ~30 builtin presets and their purpose
• Preset Structure — YAML format, predicates, optional markers
• Preset Overrides — Customizing presets via global.hydra.presets
• Cluster Overrides — Manually assign live resources to preset apps or normal apps via hydra.yaml
• Preset Activation — Activation logic, chains, debugging

Overview

A Kubernetes cluster contains many resources that are not deployed by your apps — CoreDNS, kube-proxy, flannel, metrics-server, etc. These are part of the cluster infrastructure.

Without presets, these resources would appear as untracked in hydra gitops untracked. Presets tell Hydra: "these resources belong to the cluster infrastructure."

How Presets Work

1. Each preset defines predicates (IDs or CEL expressions) that match infrastructure resources.
2. Hydra represents each enabled preset as a preset app, for example in-cluster.preset.coredns.
3. Preset anchors and matches assign resources to that preset app in the shared resource model.
4. Child resources are then attached through the same owner-reference and generic-ref logic used for normal apps.

Inspecting Presets

# Show which presets match against live cluster state
hydra gitops system prod

# Include disabled presets
hydra gitops system prod --all

Prerequisites

• Concepts: Dependency Graph
• CEL — CEL expressions are used in preset predicates