hydra gitops inspect¶

Interactively browse the reference graph for a Hydra resource using locally rendered templates combined with a live cluster connection.

Hydra loads rendered apps and live cluster inventory once per command invocation (including when you use the id picker without <id>). The TUI then navigates the reference graph in memory without reloading the cluster.

The picker status rows now come from the same resource model rows used by other cluster commands: one normalized resource ID can carry both template presence and live presence, so picker rows classify IDs directly as ok, template only, or cluster only from that shared model.

CLI help recording¶

Synopsis¶

hydra gitops inspect <cluster> [id] [flags]

Data Source¶

hydra gitops inspect renders the named cluster's applications locally and uses the live Kubernetes API for additional context during ref resolution (ConfigMap parsers, clone materialization, CEL options).

Hydra still builds the merged inspect reference graph from the shared inventory snapshot, but the template/live status shown in the picker is now read from the resource model instead of two independently queried collections.

The reference list uses transitive breadth-first expansion (max 10 levels per direction) on the merged template+live graph and shows Dist as the first list column. Unlike local inspect, cluster inspect also shows Status in the picker, header, and list; see the shared inspect documentation.

The optional <id> may refer to any resource Hydra loads from the cluster (for example a live Pod created by an operator), not only objects that appear in rendered templates.

Before running this command, validate the kubeconfig context with hydra gitops validate-current-context.

For a purely offline variant that does not require cluster connectivity, use hydra local inspect.

Shared Behavior¶

The interactive TUI layout, keyboard controls, filter popup, and navigation behavior are identical to hydra local inspect. See the shared inspect documentation (including Detail panel) for layout and keys.

Arguments¶

Argument	Description
`cluster`	Cluster name (same segment as in app ids; must not contain `.`).
`id`	Optional. Canonical Hydra resource id. If omitted, Hydra shows a picker over template and live ids (see shared inspect docs).

Flags¶

Flag	Short	Description
`--hydra-context`		Path to the Hydra context directory (or set `HYDRA_CONTEXT` env var)
`--helm-network-mode`		Helm network mode: `online`, `local`, `offline`, or `error`
`--no-cache`		Disable persistent Helm template cache and in-process Helm-related caches for this run
`--bootstrap`		Include `global.hydra.clones` rules tagged `bootstrap` when materializing clone predictions for CEL options

Examples¶

# Validate cluster first
hydra gitops validate-current-context prod

# Picker (templates + live cluster), then graph
hydra gitops inspect prod

# Browse refs for a Deployment
hydra gitops inspect prod apps/v1/Deployment/my-ns/my-app