IaC Generation with LLMs: An Error Taxonomy and A Study on Configuration Knowledge Injection

TL;DR

This paper presents a systematic approach to improve LLM-generated Infrastructure as Code by injecting structured configuration knowledge, developing an error taxonomy, and evaluating techniques that significantly enhance correctness but reveal persistent intent alignment challenges.

Contribution

It introduces a novel error taxonomy for LLM-assisted IaC generation and demonstrates that knowledge injection techniques substantially improve technical correctness in Terraform code.

Findings

Knowledge injection increased success rate from 27.1% to 62.6%.

Enhanced error analysis facilitated targeted improvements.

Intent alignment remains a challenge despite technical gains.

Abstract

Large Language Models (LLMs) currently exhibit low success rates in generating correct and intent-aligned Infrastructure as Code (IaC). This research investigated methods to improve LLM-based IaC generation, specifically for Terraform, by systematically injecting structured configuration knowledge. To facilitate this, an existing IaC-Eval benchmark was significantly enhanced with cloud emulation and automated error analysis. Additionally, a novel error taxonomy for LLM-assisted IaC code generation was developed. A series of knowledge injection techniques was implemented and evaluated, progressing from Naive Retrieval-Augmented Generation (RAG) to more sophisticated Graph RAG approaches. These included semantic enrichment of graph components and modeling inter-resource dependencies. Experimental results demonstrated that while baseline LLM performance was poor (27.1% overall success), injecting structured configuration knowledge increased technical validation success to 75.3% and overall success to 62.6%. Despite these gains in technical correctness, intent alignment plateaued, revealing a "Correctness-Congruence Gap" where LLMs can become proficient "coders" but remain limited "architects" in fulfilling nuanced user intent.