Skip to content

agntcy/dir-staging

Directory Deployment

This repository contains the deployment manifests for AGNTCY Directory project. It is designed to be used with Argo CD for GitOps-style continuous deployment.

The manifests are organized into two main sections:

  • projects/: Contains Argo CD project definitions.
  • projectapps/: Contains Argo CD application definitions.

The project will deploy the following components:

  • applications/dir - AGNTCY Directory server with storage backend (v0.5.5)
  • applications/dir-admin - AGNTCY Directory Admin CLI client (v0.5.5)
  • applications/spire* - SPIRE stack for identity and federation (with SPIFFE CSI driver)

NOTE: This is not a production-ready deployment. It is provided as-is for demonstration and testing purposes.

Latest Version: v0.5.5 - See CHANGELOG.md for what's new.

Onboarding

To onboard a new environment to Directory Public Staging Network, check the onboarding guide.

Quick Start (Development Environment)

Note

This guide sets up the development environment for local testing and development. It uses a local Kind cluster with NodePort services and simplified security. For production deployment with Ingress and TLS, see the Production Setup section below.

This guide demonstrates how to set up AGNTCY Directory project using Argo CD in Kubernetes Kind cluster.

  1. Create Kind cluster
kind create cluster --name dir-dev
  1. Install Argo CD in the cluster
# Install ArgoCD
kubectl create namespace argocd
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml

# Wait for ArgoCD to be ready
kubectl wait --namespace argocd --for=condition=available deployment --all --timeout=120s
  1. Deploy Directory via ArgoCD
# Add project
kubectl apply -f https://raw.githubusercontent.com/agntcy/dir-staging/main/projects/dir/dev/dir-dev.yaml

# Add application
kubectl apply -f https://raw.githubusercontent.com/agntcy/dir-staging/main/projectapps/dir/dev/dir-dev-projectapp.yaml
  1. Check results in ArgoCD UI
# Retrieve password
kubectl get secret argocd-initial-admin-secret -n argocd -o jsonpath="{.data.password}" | base64 -d; echo

# Port forward the ArgoCD API to localhost:8080
kubectl port-forward svc/argocd-server -n argocd 8080:443

Login to the UI at https://localhost:8080 with username admin and the password retrieved above.

Verify deployment by checking the results of CronJobs in dir-admin application.

  1. Clean up
kind delete cluster --name dir-dev

Token-based Directory Client Authentication (Dev)

In some cases, you may want to use Directory Client locally without SPIRE stack. In this case, you can use token-based authentication using SPIFFE X509 SVID tokens.

To generate a SPIFFE SVID token for authenticating local Directory Client with the Directory Server in the dev environment, follow these steps:

  1. Create a SPIFFE SVID for local Directory Client
kubectl exec spire-dir-dev-argoapp-server-0 -n dir-dev-spire -c spire-server -- \
   /opt/spire/bin/spire-server x509 mint \
   -dns dev.api.example.org \
   -spiffeID spiffe://example.org/local-client \
   -output json > spiffe-dev.json
  1. Set SPIFFE Token variable for Directory Client
# Set authentication method to token
export DIRECTORY_CLIENT_AUTH_MODE="token"
export DIRECTORY_CLIENT_SPIFFE_TOKEN="spiffe-dev.json"

# Set Directory Server address and skip TLS verification
export DIRECTORY_CLIENT_SERVER_ADDRESS="127.0.0.1:8888"
export DIRECTORY_CLIENT_TLS_SKIP_VERIFY="true"
  1. Port-forward Directory Server API
kubectl port-forward svc/dir-dir-dev-argoapp-apiserver -n dir-dev-dir 8888:8888
  1. Run Directory Client
dirctl info baeareiesad3lyuacjirp6gxudrzheltwbodtsg7ieqpox36w5j637rchwq

Production Deployment

This example configuration uses simplified settings for local Kind/Minikube testing. For production deployment, consider these enhancements:

This Example vs Production

Feature This Example (Kind) Production
SPIFFE CSI Driver ✅ Enabled (v0.5.5+) ✅ Enabled
Storage emptyDir (ephemeral) PVCs (persistent)
Deployment Strategy Recreate (default) Recreate (required with PVCs)
Credentials Hardcoded in values.yaml ExternalSecrets + Vault
Resources 250m/512Mi 500m-2000m / 1-4Gi
Ingress NodePort (local) Ingress + TLS
Rate Limits 50 RPS 500+ RPS
Trust Domain example.org your-domain.com
Read-Only FS No (emptyDir) Yes (with PVCs)

This configuration is optimized for local testing. For production, enable the optional features documented below.

Key Production Features

SPIFFE CSI Driver (v0.5.5+):

  • Enabled by default via spire.useCSIDriver: true
  • Provides synchronous workload identity injection
  • Eliminates authentication race conditions ("certificate contains no URI SAN" errors)
  • More secure than hostPath mounts in workload containers

Persistent Storage:

  • Enable PVCs for routing datastore and database (v0.5.2+)
  • Prevents data loss across pod restarts
  • See pvc.create and database.pvc.enabled in values.yaml
  • IMPORTANT: When using PVCs, set strategy.type: Recreate to prevent database lock conflicts

Secure Credential Management:

  • Use ExternalSecrets Operator with Vault instead of hardcoded secrets
  • See commented ExternalSecrets configuration in values.yaml
  • Reference: agntcy-deployment repository for production patterns

Resource Sizing:

  • Increase limits based on expected load (CPU: 500m-2000m, Memory: 1-4Gi)
  • Monitor and adjust after observing production traffic

Ingress & TLS:

  • Configure Ingress for external access
  • Use cert-manager with Let's Encrypt for production certificates
  • Enable SSL passthrough for DIR API (SPIFFE mTLS)

Minikube Production Simulation

If you wish to test production-like setup locally with Ingress and TLS, follow the steps below using Minikube.

Note

We are using Minikube to simulate production setup, as it supports Ingress and TLS out of the box. Steps below marked as (local) are optional and intended for local testing purposes only.

Caution

It is not recommended to deploy both dev and prod environments in the same cluster, as they may conflict with each other.

View Production Setup
  1. Create Minikube cluster
minikube start -p dir-prod
  1. (local) Enable Ingress and DNS addons in Minikube

The deployment uses *.test domain for Ingress resources.

For local testing purposes, Minikube Ingress controller is required to route traffic to our Ingress resources.

Otherwise, if you are deploying to a cloud provider with its own Ingress controller, make sure that it supports SSL Passthrough.

# Enable Ingress and Ingress-DNS addons
minikube addons enable ingress -p dir-prod
minikube addons enable ingress-dns -p dir-prod

# Patch Ingress controller to enable SSL Passthrough
kubectl patch deployment -n ingress-nginx ingress-nginx-controller --type='json' \
-p='[{"op": "add", "path": "/spec/template/spec/containers/0/args/-", "value":"--enable-ssl-passthrough"}]'
  1. (local) Enable Local DNS inside Minikube

The deployment uses *.test domain for Ingress resources.

For local testing purposes, we need to configure DNS resolution inside Minikube cluster to resolve *.test domain to Minikube IP address using minikube/ingress-dns guide.

Otherwise, if you are deploying to a cloud provider with its own Ingress controller, you can skip this step.

# Get Minikube IP
minikube ip -p dir-prod

# Add DNS resolver entry for `*.test` domain
# Follow guide at: https://minikube.sigs.k8s.io/docs/handbook/addons/ingress-dns

# Update CoreDNS ConfigMap to forward `test` domain to Minikube IP
kubectl edit configmap coredns -n kube-system
  1. (local) Install CertManager with Self-Signed Issuer

The deployment uses CertManager letsencrypt issuer to issue TLS certificates for Ingress resources.

For local testing purposes, we will create a self-signed root CA certificate and configure CertManager to use it as letsencrypt issuer.

Otherwise, if you are deploying to a cloud provider with its own CertManager, you can skip this step, but ensure that letsencrypt issuer is available.

# Install Cert-Manager
kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.19.1/cert-manager.yaml

# Wait for Cert-Manager to be ready
kubectl wait --namespace cert-manager --for=condition=available deployment --all --timeout=120s

# Create Self-Signed Issuer and Root CA Certificate
kubectl apply -f - <<EOF
apiVersion: cert-manager.io/v1
kind: ClusterIssuer
metadata:
   name: selfsigned-issuer
spec:
   selfSigned: {}
---
apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
   name: my-selfsigned-ca
   namespace: cert-manager
spec:
   isCA: true
   commonName: my-selfsigned-ca
   secretName: root-secret
   privateKey:
      algorithm: ECDSA
      size: 256
   issuerRef:
      name: selfsigned-issuer
      kind: ClusterIssuer
      group: cert-manager.io
---
apiVersion: cert-manager.io/v1
kind: ClusterIssuer
metadata:
   name: letsencrypt
spec:
   ca:
      secretName: root-secret
EOF
  1. Install Argo CD in the cluster
# Install ArgoCD
kubectl create namespace argocd
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml

# Wait for ArgoCD to be ready
kubectl wait --namespace argocd --for=condition=available deployment --all --timeout=120s
  1. Deploy Directory via ArgoCD
# Add project
kubectl apply -f https://raw.githubusercontent.com/agntcy/dir-staging/main/projects/dir/prod/dir-prod.yaml

# Add application
kubectl apply -f https://raw.githubusercontent.com/agntcy/dir-staging/main/projectapps/dir/prod/dir-prod-projectapp.yaml
  1. Check results in ArgoCD UI:
# Retrieve password
kubectl get secret argocd-initial-admin-secret -n argocd -o jsonpath="{.data.password}" | base64 -d; echo

# Port forward the ArgoCD API to localhost:8080
kubectl port-forward svc/argocd-server -n argocd 8080:443

Login to the UI at https://localhost:8080 with username admin and the password retrieved above.

Verify deployment by checking the results of CronJobs in dir-admin application.

  1. Clean up
minikube delete -p dir-prod

Token-based Directory Client authentication

To generate a SPIFFE SVID token for authenticating local Directory Client with the Directory Server, follow these steps:

  1. Create a SPIFFE SVID for local Directory Client
kubectl exec spire-dir-prod-argoapp-server-0 -n dir-prod-spire -c spire-server -- \
   /opt/spire/bin/spire-server x509 mint \
   -dns prod.api.example.org \
   -spiffeID spiffe://example.org/local-client \
   -output json > spiffe-prod.json
  1. Set SPIFFE Token variable for Directory Client
# Set authentication method to token
export DIRECTORY_CLIENT_AUTH_MODE="token"
export DIRECTORY_CLIENT_SPIFFE_TOKEN="spiffe-prod.json"

# Set Directory Server address (via Ingress)
export DIRECTORY_CLIENT_SERVER_ADDRESS="prod.api.example.org:443"

# Or, set Directory Server address and skip TLS verification (via port-forwarding)
export DIRECTORY_CLIENT_SERVER_ADDRESS="127.0.0.1:8888"
export DIRECTORY_CLIENT_TLS_SKIP_VERIFY="true"
  1. Port-forward Directory Server API
kubectl port-forward svc/dir-dir-prod-argoapp-apiserver -n dir-prod-dir 8888:8888
  1. Run Directory Client
dirctl info baeareiesad3lyuacjirp6gxudrzheltwbodtsg7ieqpox36w5j637rchwq

Copyright Notice

Copyright Notice and License

Distributed under Apache 2.0 License. See LICENSE for more information. Copyright AGNTCY Contributors (https://github.com/agntcy)

About

Agent Directory Service Staging GitOps

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •