Command Overview

All Dorgu CLI commands at a glance

​

Command tree

dorgu
├── health                  Show cluster health summary
├── incidents               View detected incidents
│   ├── list                List incidents
│   └── describe <name>     Show incident details
├── remediation             Review and act on proposed fixes
│   ├── list                List remediation proposals
│   ├── diff <name>         Show the proposed change
│   ├── approve <name>      Approve a pending remediation and heal the workload
│   ├── reject <name>       Reject a remediation proposal, with a reason
│   └── heal <name>         Apply an approved remediation's ticket to the workload
├── report                  Local summary of how the loop is doing on this cluster
├── persona                 ApplicationPersona CRDs
│   ├── import              Create personas from Deployments already running
│   ├── list                List personas across namespaces
│   └── status <name>       Check persona status
├── cluster                 ClusterPersona CRDs
│   └── status [name]       View cluster status
├── version                 Show version info
└── completion              Generate a shell completion script

dorgu incidents has only list and describe. There is no history subcommand.

CLI v0.13.0 removed every command that was not part of the self-healing loop: generate, init, config, persona generate, persona apply, cluster init, cluster setup, cluster info, platform serve, watch, sync, health --watch, and the global --config flag. The CLI migration guide lists what to use instead for each one.

​

Command reference

The command tree above is a plain listing. Use these links for the full flag reference on each command group.

Command	Reference	What it does
dorgu health	health	Cluster health summary
dorgu incidents	incidents	List and inspect detected incidents
dorgu remediation	remediation	Review, approve, reject, and apply proposed fixes
dorgu report	report	Local summary of safe heals, approvals, refusals and timings (text or JSON)
dorgu persona	persona	Import and inspect ApplicationPersona CRDs
dorgu cluster	cluster	Inspect the ClusterPersona the operator maintains
dorgu version	version	Show version info
dorgu completion	dorgu completion --help	Shell completion for bash, zsh, fish, and PowerShell

​

Global flags

These flags apply to all commands:

Flag	Description	Default
--context	kube-context for every kubectl call the command makes	the kubeconfig’s current-context
--json	Output as JSON	false
--no-color	Disable colored output	false

Cluster commands also take --kubeconfig. Both it and --context are applied to every kubectl call a command makes, so a command never reads one cluster and writes another. heal’s production-context guard checks the --context value when you set it, and the kubeconfig’s current-context otherwise.

​

Prerequisites

Every command that touches a cluster shells out to kubectl, so it must be on your PATH, and either your current context or --context must point at the target cluster. Without it those commands fail with kubectl not found in PATH. Only version and completion work with no cluster at all.

Set DORGU_DEBUG=1 to see kubectl’s raw output in error messages. By default the CLI strips klog noise (the memcache.go:265 "Unhandled Error" lines) so the message that says what went wrong is not buried.

​

Exit codes

Code	Meaning
0	Success
1	Error: the command could not run. An unreachable cluster, a missing kubeconfig, no kubectl, a failed API call.
2	dorgu health --exit-code only: active critical incidents
3	dorgu health --exit-code only: health could not be judged
4	dorgu remediation approve / heal only: declined by design. Dorgu understood the plan and chose not to write, because something else owns the workload or because the operator’s gate refused the approval. Nothing failed and nothing was applied.

Codes 2 and 3 are opt-in and specific to dorgu health. An unknown subcommand is an error with a non-zero exit, not a help dump.

4 is a decision, not a failure. A wrapper that treats every non-zero exit as breakage will report a successful refusal as an error. Accept 4, or use --no-heal, which records the decision without a workload patch and exits 0. See the ownership model.

​

Getting help

# General help
dorgu --help

# Command-specific help
dorgu remediation approve --help
dorgu persona import --help