Project #2 · Infrastructure Automation
ClusterCollie AI Playbook (Ansible)
Describe an infrastructure task in plain English. Get back a validated, syntax-checked Nutanix Ansible playbook — grounded in the real nutanix.ncp collection, not hallucinated YAML.
What It Does
Nutanix publishes an official Ansible collection (nutanix.ncp) covering Prism Central, Foundation, Foundation Central, Karbon, and NDB, with modules like ntnx_vms_v2, ntnx_clusters_v2, and ntnx_foundation. Writing correct playbooks against this collection still requires knowing exact module names, parameter schemas, and API version differences (v1 vs. v2 resources). ClusterCollie AI Playbook removes that friction: describe what you want — for example, "provision a 3-node VM cluster with 16 vCPU and 64GB RAM each, tagged for the finance team" — and the tool generates a ready-to-review Ansible playbook using the real module schema.
How It Works
Enter a natural-language infrastructure request via CLI or a simple web form — node counts, vCPU/RAM sizing, network tags, categories, and target cluster.
The tool retrieves relevant example playbooks and module documentation from the official
nutanix.ncp collection (via few-shot prompting or retrieval-augmented generation) so the LLM drafts YAML against real, current module names and parameters instead of inventing syntax.Before any output is shown, the tool automatically runs
ansible-playbook --syntax-check against the generated file. If it fails, the tool retries generation with the error fed back into the prompt, rather than handing you broken YAML.The validated playbook is presented with a plain-English summary of what it will do, so you can review scope and blast radius before running it against real infrastructure.
Where supported by the target modules, the tool can suggest a
--check mode dry run as an additional guardrail prior to live execution.Grounding Source
The generator's reference set is built from the official nutanix/nutanix.ansible GitHub repository, including its examples/ directories for Prism Central, Foundation, and Foundation Central playbooks. Anchoring generation to real examples and the actual module list is what separates this tool from a generic "ask an LLM for YAML" approach.
Operational Guardrails
- Mandatory syntax-check pass before any playbook is presented as final
- Clear labeling of AI-generated content per the site-wide AI Disclosure
- Recommended read-only/staging validation before production runs
- No embedded credentials — playbooks reference Ansible Vault-style variable placeholders
Why It Matters
This project pairs practical DevOps tooling with applied AI — skills that map directly to Infrastructure Automation Engineer roles that reference Nutanix explicitly. It demonstrates not just "can an LLM write YAML" but the harder, more valuable skill: constraining an LLM's output to match a real, versioned API surface and building the validation guardrails that make AI-generated infrastructure code safe to review rather than a liability.