The argocd-resource-tracker CLI provides command-line parameters to control its behavior. This document describes all available commands and flags.
argocd-resource-tracker analyze [flags]Analyzes resource relationships and dependencies for ArgoCD applications. Can process a single application or all applications in a namespace. Outputs resource.inclusions YAML that can be used to configure ArgoCD's argocd-cm ConfigMap.
| Flag | Short | Default | Description |
|---|---|---|---|
--app |
Application name (required for single app analysis). Supports namespace/name syntax (e.g., argocd/my-app) |
||
--app-namespace |
-N |
argocd |
Namespace where the application is located |
--all-apps |
false |
Analyze all applications in the namespace | |
--strategy |
dynamic |
Analysis strategy: dynamic (OwnerRef walking) or graph (Cyphernetes) |
|
--namespace |
-n |
argocd |
ArgoCD namespace (where ArgoCD is installed) |
--kubeconfig |
Path to kubeconfig file. If not provided, uses in-cluster config or ~/.kube/config |
||
--repo-server |
Repo server address. If empty, CLI will port-forward to argocd-repo-server service |
||
--repo-server-plaintext |
false |
Use unencrypted HTTP connection to repo-server (instead of TLS) | |
--repo-server-strict-tls |
false |
Enable strict TLS validation for repo-server connection | |
--repo-server-timeout-seconds |
60 |
Timeout in seconds for repo-server RPC calls | |
--loglevel |
info |
Log level: trace, debug, info, warn, or error |
The dynamic strategy uses the Kubernetes discovery client to list all API resource kinds (including CRDs) on the cluster and builds an in-memory parent→child cache from owner references. For each resource returned by the Argo CD repo-server (the application’s immediate children), it looks up that kind in the cache, then recursively looks up each child kind until the full dependency tree is built.
The Graph strategy uses the Cyphernetes library to run graph based relationship queries against the cluster. For each resource returned by the Argo CD repo-server (the application’s immediate children), it runs a Cyphernetes query to find that resource’s children, then repeats this process recursively for each child until the full dependency tree is built.