Skip to main content
Agent installation is only required for Agent mode. If your source and destination are cloud services reachable over the internet (Snowflake, BigQuery, RDS with a public endpoint, SaaS APIs), you can use Hosted mode and skip this page entirely. See Execution Modes to decide which fits your setup.
The Zwiron Agent is a lightweight binary that runs inside your network. It:
  • Connects outbound to Zwiron — no inbound ports required
  • Reads data from your sources and writes to your destinations
  • Decrypts Agent-mode credentials locally for syncs (RSA key unique to the agent)
  • Runs as a background system service (auto-starts on reboot)

Prerequisites


The fastest way to get started on any Linux server or Mac:
Then register with your token:
The installer will:
  1. Detect your OS and CPU architecture
  2. Download the latest agent binary
  3. Install it to /usr/local/bin/zwiron-agent
  4. Register it as a systemd service (Linux) or launchd service (macOS)
  5. Start the service automatically
Verify it’s running:

Option 2 — Docker

Run the agent as a Docker container. Useful when you want container-level isolation or you’re already running Docker on your host.
The agent stores its RSA keys in /data inside the container. Mount a volume if you want keys to persist across container restarts.

Option 3 — Kubernetes

Deploy the agent as a Deployment in your cluster. This is recommended for production environments.
Create the token secret first:

Agent commands


How the agent connects

The agent uses a persistent outbound gRPC stream to Zwiron. This means:
  • No inbound firewall rules needed
  • Works behind NAT, corporate firewalls, and VPNs
  • If the connection drops, the agent reconnects automatically with exponential backoff
All communication is encrypted with TLS. Agent-mode database credentials are encrypted with an RSA key unique to your agent — Zwiron stores ciphertext and cannot read plaintext without the agent’s private key.

Managing multiple agents

You can install multiple agents — for example, one per environment (staging, production) or one per network segment. Each agent shows up independently in your dashboard under Settings → Agents. When creating a connection, you select which agent should handle that connection. This lets you route traffic precisely — for example, keeping production data on its own agent.

Troubleshooting

Check that the agent service is running: sudo zwiron-agent statusIf it’s stopped: sudo systemctl start zwiron-agent (Linux) or sudo launchctl start zwiron-agent (macOS)Check logs for errors: sudo zwiron-agent logs
The agent runs inside your network and connects to your databases directly. Ensure:
  • The database host is reachable from the machine where the agent is running
  • The database port is not blocked by a local firewall
  • The database user has the required permissions (see the connector-specific guide for your database)
The agent needs outbound access on port 443 to agent.zwiron.com. Check your egress firewall rules and proxy settings.
Last modified on July 13, 2026