Deploying Tailscale For Zero Trust Vpn

Deploying Tailscale for Zero Trust VPN

Overview

Tailscale is a zero trust mesh VPN built on WireGuard that creates encrypted peer-to-peer connections between devices without requiring traditional VPN servers or complex network configuration. Every connection in a Tailscale network (tailnet) is end-to-end encrypted using WireGuard's Noise protocol framework with Curve25519 key exchange. Tailscale implements zero trust networking by authenticating every connection request through identity providers, enforcing granular Access Control Lists (ACLs), and supporting features like exit nodes, subnet routers, MagicDNS, and Tailscale SSH. For organizations preferring self-hosted infrastructure, Headscale provides an open-source implementation of the Tailscale control server.

When to Use

• When deploying or configuring deploying tailscale for zero trust vpn capabilities in your environment
• When establishing security controls aligned to compliance requirements
• When building or improving security architecture for this domain
• When conducting security assessments that require this implementation

Common Misconfigurations & Verification

• Default-allow ACL left in place: a new tailnet ships with {"action":"accept","src":["*"],"dst":["*:*"]}. If that wildcard rule remains, every node reaches every other node on every port, zero trust in name only. Replace it with scoped tag-based src/dst rules and confirm the policy no longer contains "*:*".
• Subnet router exposes more than intended: --advertise-routes=10.0.0.0/8 plus --accept-routes can hand clients the entire datacenter. Advertise the narrowest CIDR and gate it behind autoApprovers/ACL dst tags.
• Key expiry disabled on nodes: --authkey servers with key expiry off never force re-auth; pair with Tailnet Lock so unauthorized nodes cannot join.
• Tailscale SSH using accept not check: accept grants a shell with no re-auth or session recording; use check for production tags, and confirm exit nodes aren't unintentionally in autoApprovers.

Verify: from a node that should be denied, run tailscale ping <peer> and attempt the service port, it must fail, and tailscale status should show no route. Review the admin console audit/network-flow logs to confirm blocked attempts are recorded, and run tailscale up from an unsigned node to confirm Tailnet Lock rejects it.

Prerequisites

• Identity provider (Okta, Azure AD, Google Workspace, GitHub, or OIDC-compatible)
• Devices running supported OS (Linux, Windows, macOS, iOS, Android, FreeBSD)
• Administrative access to configure DNS and firewall rules
• Understanding of WireGuard protocol fundamentals
• Network planning documentation for subnet routing requirements

Architecture

                    Tailscale Coordination Server
                    (or self-hosted Headscale)
                           |
                    Key Distribution
                    & NAT Traversal
                           |
         +-----------------+-----------------+
         |                 |                 |
    +----+----+      +----+----+      +----+----+
    | Node A  |<---->| Node B  |<---->| Node C  |
    | (Linux) |      | (macOS) |      |(Windows)|
    +---------+      +---------+      +---------+
    WireGuard         WireGuard        WireGuard
    Encrypted         Encrypted        Encrypted
    P2P Tunnel        P2P Tunnel       P2P Tunnel

    Each node connects directly to every other node.
    DERP relay servers used only when direct P2P fails.

Installation and Setup

Linux Installation

# Add Tailscale repository and install
curl -fsSL https://tailscale.com/install.sh | sh

# Start Tailscale and authenticate
sudo tailscale up

# Check connection status
tailscale status

# View assigned IP address
tailscale ip -4
tailscale ip -6

Windows / macOS Installation

# Windows: Download from https://tailscale.com/download/windows
# macOS: Install via Homebrew
brew install --cask tailscale

# Or download from https://tailscale.com/download/mac

Docker Deployment

# docker-compose.yml for Tailscale sidecar
version: '3.8'
services:
  tailscale:
    image: tailscale/tailscale:latest
    container_name: tailscale
    hostname: my-service
    environment:
      - TS_AUTHKEY=tskey-auth-xxxxx  # Pre-auth key
      - TS_STATE_DIR=/var/lib/tailscale
      - TS_EXTRA_ARGS=--advertise-tags=tag:container
    volumes:
      - tailscale-state:/var/lib/tailscale
      - /dev/net/tun:/dev/net/tun
    cap_add:
      - net_admin
      - sys_module
    restart: unless-stopped

volumes:
  tailscale-state:

Kubernetes Deployment

# Tailscale operator for Kubernetes
apiVersion: v1
kind: Secret
metadata:
  name: tailscale-auth
  namespace: tailscale
type: Opaque
stringData:
  TS_AUTHKEY: "tskey-auth-xxxxx"
---
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: tailscale
  namespace: tailscale
spec:
  selector:
    matchLabels:
      app: tailscale
  template:
    metadata:
      labels:
        app: tailscale
    spec:
      containers:
      - name: tailscale
        image: tailscale/tailscale:latest
        env:
        - name: TS_AUTHKEY
          valueFrom:
            secretKeyRef:
              name: tailscale-auth
              key: TS_AUTHKEY
        - name: TS_KUBE_SECRET
          value: tailscale-state
        - name: TS_USERSPACE
          value: "true"
        securityContext:
          capabilities:
            add: ["NET_ADMIN"]

Access Control Lists (ACLs)

Tailscale ACLs define who can access what within your tailnet using a declarative JSON format. The default policy is deny-all, making it zero trust by design.

{
  "acls": [
    // Engineering team can access development servers
    {
      "action": "accept",
      "src": ["group:engineering"],
      "dst": ["tag:dev-server:*"]
    },
    // SRE team can access production infrastructure
    {
      "action": "accept",
      "src": ["group:sre"],
      "dst": ["tag:production:22,443,8080"]
    },
    // Database access restricted to backend services
    {
      "action": "accept",
      "src": ["tag:backend"],
      "dst": ["tag:database:5432,3306,27017"]
    },
    // All employees can access internal tools
    {
      "action": "accept",
      "src": ["group:employees"],
      "dst": ["tag:internal-tools:443"]
    }
  ],

  "groups": {
    "group:engineering": ["user@company.com", "dev@company.com"],
    "group:sre": ["sre@company.com", "oncall@company.com"],
    "group:employees": ["autogroup:members"]
  },

  "tagOwners": {
    "tag:dev-server": ["group:engineering"],
    "tag:production": ["group:sre"],
    "tag:backend": ["group:sre"],
    "tag:database": ["group:sre"],
    "tag:internal-tools": ["group:sre"],
    "tag:container": ["group:sre"]
  },

  "ssh": [
    {
      "action": "check",
      "src": ["group:sre"],
      "dst": ["tag:production"],
      "users": ["root", "admin"]
    },
    {
      "action": "accept",
      "src": ["group:engineering"],
      "dst": ["tag:dev-server"],
      "users": ["autogroup:nonroot"]
    }
  ],

  "nodeAttrs": [
    {
      "target": ["autogroup:members"],
      "attr": ["funnel:deny"]
    }
  ]
}

Exit Nodes and Subnet Routing

Configure Exit Node

# On the exit node machine
sudo tailscale up --advertise-exit-node

# On the client machine, use the exit node
sudo tailscale up --exit-node=<exit-node-ip>

# Verify exit node routing
curl ifconfig.me  # Should show exit node's public IP

Subnet Router Configuration

# Advertise local subnets through Tailscale
sudo tailscale up --advertise-routes=10.0.0.0/24,192.168.1.0/24

# Enable IP forwarding on Linux
echo 'net.ipv4.ip_forward = 1' | sudo tee -a /etc/sysctl.conf
echo 'net.ipv6.conf.all.forwarding = 1' | sudo tee -a /etc/sysctl.conf
sudo sysctl -p

# Accept routes on client
sudo tailscale up --accept-routes

Tailscale SSH (Zero Trust SSH)

Tailscale SSH replaces traditional SSH key management with identity-based access.

# Enable Tailscale SSH on a server
sudo tailscale up --ssh

# Connect using Tailscale SSH (no SSH keys needed)
ssh user@hostname  # Authenticates via Tailscale identity

# Session recording (audit logging)
# Configure in ACL policy:
# "ssh": [{"action": "check", "src": [...], "dst": [...], "users": [...]}]
# "check" action requires re-authentication and records sessions

MagicDNS Configuration

# MagicDNS is enabled by default in new tailnets
# Access devices by hostname instead of IP
ping my-server  # Resolves via MagicDNS

# Custom DNS configuration via admin console
# Split DNS: route specific domains to internal DNS servers
# Global nameservers: override default DNS resolution

Self-Hosted with Headscale

# Install Headscale (open-source Tailscale control server)
wget https://github.com/juanfont/headscale/releases/latest/download/headscale_linux_amd64
chmod +x headscale_linux_amd64
sudo mv headscale_linux_amd64 /usr/local/bin/headscale

# Create configuration
sudo mkdir -p /etc/headscale
sudo headscale generate config > /etc/headscale/config.yaml

# Edit config for your environment
# Key settings:
#   server_url: https://headscale.example.com
#   listen_addr: 0.0.0.0:8080
#   private_key_path: /etc/headscale/private.key
#   db_type: sqlite3
#   db_path: /var/lib/headscale/db.sqlite

# Start Headscale
sudo headscale serve

# Create user and pre-auth key
headscale users create myorg
headscale preauthkeys create --user myorg --reusable --expiration 24h

# Connect Tailscale client to Headscale
tailscale up --login-server https://headscale.example.com

Security Hardening

Key Expiry and Rotation

# Set key expiry in admin console (default: 180 days)
# Force re-authentication periodically

# Disable key expiry for servers (use auth keys instead)
sudo tailscale up --authkey=tskey-auth-xxxxx

# Pre-auth keys for automated deployment
# Create ephemeral, single-use keys for CI/CD

Device Authorization

{
  "nodeAttrs": [
    {
      "target": ["autogroup:members"],
      "attr": [
        "mullvad:deny",
        "funnel:deny"
      ]
    }
  ],
  "autoApprovers": {
    "routes": {
      "10.0.0.0/24": ["group:sre"],
      "192.168.0.0/16": ["group:sre"]
    },
    "exitNode": ["group:sre"]
  }
}

Network Lock (Tailnet Lock)

# Initialize network lock with signing keys
tailscale lock init

# Add trusted signing keys
tailscale lock add nodekey:xxxxx

# All new nodes require signing before joining
# Prevents unauthorized nodes from joining the tailnet

Monitoring and Observability

# View network status
tailscale status --json | jq '.Peer | to_entries[] | {name: .value.HostName, online: .value.Online, os: .value.OS}'

# Check connection quality
tailscale ping <peer-ip>

# View network map
tailscale netcheck

# Audit logs available in Tailscale admin console
# Integration with SIEM via webhook or API

Integration Patterns

Service Mesh Integration

# Tailscale as sidecar for service-to-service communication
# Each service gets a Tailscale identity
# ACLs enforce service-to-service access policies

# Example: API service can only reach database service
# ACL: tag:api -> tag:database:5432

CI/CD Pipeline Integration

# Use ephemeral auth keys in CI/CD
export TS_AUTHKEY=tskey-auth-xxxxx-ephemeral
tailscale up --authkey=$TS_AUTHKEY --hostname=ci-runner-$CI_JOB_ID

# Access internal resources during build/deploy
# Node automatically removed when container stops

References

• Tailscale Documentation
• How Tailscale Works
• Tailscale ACL Documentation
• Headscale - Open Source Control Server
• WireGuard Protocol
• Tailscale SSH