How martimy/network-automation-and-observability works
Project type and purpose
• This is a network-automation lab environment demonstrating multi-vendor automation and observability in a leaf-spine data-center fabric.
• It uses Containerlab to spin up six router containers (two Nokia SR Linux spines and four NVIDIA Cumulus Linux leaves) and shows how to drive them via gNMI/YANG and NVUE REST API.
File structure
• Top-level tutorials (Markdown) describe lab setup and the network-automation tutorial.
• Configuration files:
• topology.clab.yml – container topology definition
• inventory.yaml, ansible.cfg, requirements.yaml – Ansible inventory and dependencies
• extract.sh – helper script for extracting or preparing artifacts
• No separate “src” or “frontend/backend” directories—this is a lab playground focused on automation scripts and platform interactions.
Key insights from the README
• Lab topology: two spines + four leaves connected in a leaf/spine fabric.
• Management network: 192.168.121.0/24 across all containers.
• Automation interfaces:
• Spines: Nokia SR Linux via gNMI (YANG) driven by tools like gnmic.
• Leaves: Cumulus Linux via NVUE REST API driven by curl or Ansible.
• Future phases: unified automation framework, CI/CD, observability—though Phase 1 is direct device interaction.
Main components
a. Containerlab Fabric
• Spine1 (Nokia SR Linux)
• Spine2 (Nokia SR Linux)
• Leaf1–Leaf4 (NVIDIA Cumulus Linux)
• Interconnect lines showing leaf-to-spine links in a classic leaf/spine mesh.
b. Management Network
• A network cloud labeled “Mgmt Network: 192.168.121.0/24”
c. Automation Layer
• Ansible (inventory.yaml, ansible.cfg, requirements.yaml)
• gnmic (gNMI CLI)
• curl (NVUE REST calls)
• extract.sh (helper script)
d. External References
Architectural patterns and design principles
• Physical-logical separation: data-plane (leaf/spine fabric) vs. control/management-plane (Mgmt Network + automation).
• Multi-vendor abstraction: one toolchain (Ansible + gNMI + REST) driving two different NOS.
• Container-based virtualization for reproducibility and isolation.