Quick Start

Go from zero to a running container in five steps. Odysseus is a fully managed platform -- the control plane is hosted for you. All you need is a Linux server with Docker installed.

Estimated time: Under 5 minutes from sign-up to your first running deployment.

Step 1: Sign Up

Create your account at odysseus.delta-telematics.dev. The guided onboarding wizard walks you through initial setup:

  1. Enter your email and create a password (or sign in with SSO).
  2. Name your organization and invite team members (optional -- you can do this later).
  3. Choose your deployment path:
    • Single server -- One node for development, staging, or small workloads.
    • Multi-node cluster -- Multiple servers for production workloads with scheduling across nodes.
    • Edge -- Lightweight nodes at edge locations connected back to the control plane.

After completing the wizard, you land on the dashboard with your unique install command ready to copy.

Step 2: Connect Your First Node

Run the one-liner install script shown in your dashboard on any Linux server. The script is unique to your account and contains an enrollment token.

Prerequisites

  • Operating system: Ubuntu 20.04+, Debian 11+, or RHEL/CentOS 8+
  • Docker: Docker Engine 20.10 or later installed and running
  • Network: Outbound HTTPS (port 443) to odysseus.delta-telematics.dev
  • Network: Outbound UDP port 51820 for WireGuard VPN tunnel
  • Permissions: Root or sudo access (for the initial install only)

Network & Firewall Requirements

After enrollment, the agent communicates with the control plane over an encrypted WireGuard VPN tunnel. The following ports must be open on the WireGuard interface (wg0) for the node to function correctly:

Port Protocol Purpose
8081 TCP Container API — control plane sends deployment commands to the agent
9501 TCP Metrics and health endpoint — used by monitoring and health checks
8300 TCP Consul RPC — catalog synchronization between node and control plane
8301 TCP + UDP Consul gossip — cluster membership and failure detection
Important: These ports only need to be open on the WireGuard interface (wg0), not on the public internet. Traffic flows exclusively through the encrypted VPN tunnel. If your server has a restrictive firewall, open these ports with:
sudo iptables -I INPUT -i wg0 -p tcp --dport 8081 -j ACCEPT
sudo iptables -I INPUT -i wg0 -p tcp --dport 9501 -j ACCEPT
sudo iptables -I INPUT -i wg0 -p tcp --dport 8300 -j ACCEPT
sudo iptables -I INPUT -i wg0 -p tcp --dport 8301 -j ACCEPT
sudo iptables -I INPUT -i wg0 -p udp --dport 8301 -j ACCEPT

The install script checks for these ports automatically and provides instructions if any are blocked.

Install the Agent

Copy the command from the dashboard and run it on your server:

curl -fsSL https://odysseus.delta-telematics.dev/install | sudo bash -s -- --token YOUR_ENROLLMENT_TOKEN

The install script performs three actions:

  1. Installs the Odysseus agent -- a lightweight binary that manages containers on this node.
  2. Establishes an encrypted WireGuard VPN tunnel to the control plane for secure communication.
  3. Registers the node with your account and reports its capacity (CPU, memory, disk).
Instant feedback: Your node appears in the dashboard within seconds of running the install script. The status indicator turns green once the agent is healthy and ready to accept workloads.
Nodes view showing connected nodes with health status, VPN, scheduling, CPU and memory metrics

Step 3: Deploy Your First Container

Deploy your first container using the dashboard.

Dashboard

Click New Deployment in the dashboard. Fill in the form:

  • Name: hello-web
  • Image: nginx:alpine
  • Replicas: 2
  • Port: 80

Click Deploy. Odysseus validates the configuration, schedules containers across your nodes, pulls the image, and starts the containers.

Deployment configuration form with networking, volumes, resources, health checks, and autoscaling options
Interactive tour: First-time users see an interactive walkthrough in the dashboard that highlights each section. You can replay the tour at any time from Settings > Help > Replay Tour.
Or ask Athena: "Deploy nginx with 2 replicas"

Step 4: Scale and Monitor

Your deployment is running. Now scale it and observe the metrics.

Manual Scaling

Scale from the dashboard by editing the deployment and adjusting the replica count.

Autoscaling

Enable autoscaling from the deployment settings to let Odysseus manage replica counts automatically based on CPU utilization. Set a target CPU percentage and min/max replica bounds.

Prometheus metrics drive scaling decisions. The autoscaler evaluates metrics every 15 seconds and adjusts replicas to keep CPU utilization near your target.

Or ask Athena: "Scale my-app to 5 replicas"

Monitoring

The dashboard displays real-time metrics for every deployment:

  • CPU and memory usage per container
  • Request rates and error rates
  • Container health status and restart counts
  • Autoscaler decisions and scaling history

Step 5: Explore Advanced Features

You have a running, scalable deployment. Here is what to explore next:

Feature What It Does Try It
Canary Deployments Roll out a new image version to a percentage of traffic, monitor health, and promote or roll back. Dashboard > Deployment > Canary Release
Athena AI Ask questions and run operations in natural language from the dashboard chat panel. Open the Athena panel and type: "Show me the status of hello-web"
CVE Scanning Scan images for known vulnerabilities before deploying them. Dashboard > Security > CVE Scanner
Secrets Store sensitive values and inject them into containers securely. Dashboard > Secrets > Create Secret
RBAC Invite team members with scoped roles: Admin, Operator, Developer, or Read-only. Dashboard > Settings > Team Members
Audit Log View a tamper-evident record of every action taken on the platform. Dashboard > Audit Log

Next Steps

  • Read the User Guide for in-depth coverage of every feature.
  • Review Security to understand the platform's security model.
  • Check Integrations to connect Odysseus with your CI/CD pipeline and monitoring stack.