Skip to content

Glossary

Quick reference for terms used throughout Agentic InfraOps documentation.

Microsoft’s cloud-based identity and access management service, now branded as Microsoft Entra ID. Used for authentication and RBAC across Azure resources. SQL databases in this project require AAD-only authentication (no SQL auth).

🔗 External: Microsoft Entra ID

A document that captures an important architectural decision along with its context and consequences. Used to record “why” decisions were made for future reference.

📁 Output: agent-output/{project}/03-des-adr-*.md, 07-ab-adr-*.md

A specialized AI assistant defined in .github/agents/ that focuses on specific workflow steps. Invoked via Ctrl+Shift+A. This project includes top-level agents (including two Conductor variants, a Governance agent, and a Context Optimizer) plus subagents.

📁 See: .github/agents/

The methodology of using coordinated AI agents and skills to transform requirements into deploy-ready Azure infrastructure. Combines GitHub Copilot with custom agents and reusable skills.

Microsoft’s official library of pre-built, tested IaC modules that follow Azure best practices. Available for both Bicep (br/public:avm/res/) and Terraform (registry.terraform.io/Azure/avm-res-*/azurerm). Using AVM modules ensures policy compliance and reduces custom code.

🔗 External: Azure Verified Modules Registry

AVM-TF (Azure Verified Modules for Terraform)

Section titled “AVM-TF (Azure Verified Modules for Terraform)”

The Terraform variant of Azure Verified Modules, published to the Terraform Registry under the Azure namespace. Module sources follow the pattern Azure/avm-res-<provider>-<resource>/azurerm.

🔗 External: AVM-TF on Terraform Registry

Managed Kubernetes container orchestration service on Azure. Simplifies deploying, managing, and scaling containerised applications.

🔗 External: AKS Documentation

A set of defined rules and protocols that allows software components to communicate. In this project, agents interact with Azure and GitHub APIs via MCP servers.

Azure’s deployment and management layer. All Azure resource operations go through ARM. Bicep compiles to ARM templates (JSON). The Azure MCP server queries ARM directly.

🔗 External: ARM Overview

Azure’s domain-specific language (DSL) for deploying Azure resources declaratively. Compiles to ARM templates but with cleaner syntax and better tooling support.

🔗 External: Bicep Documentation

Static analysis tool that checks Bicep files for best practices, security issues, and common mistakes. Run with bicep lint main.bicep or automatically via VS Code extension.

Microsoft’s methodology for cloud adoption, including naming conventions, governance, and landing zone architecture. This project follows CAF naming prefixes (e.g. rg- for resource groups, vnet- for virtual networks).

🔗 External: Azure CAF

A distributed network of servers that caches and delivers content from edge locations closest to users. Azure CDN / Azure Front Door accelerate static asset delivery.

Adversarial review agent that challenges requirements, architecture assessments, and implementation plans. Finds untested assumptions, governance gaps, WAF blind spots, and architectural weaknesses. Returns structured JSON findings with severity ratings. Auto-invoked by the Conductor after Steps 1, 2, and 4.

📁 See: .github/agents/10-challenger.agent.md

The conversational interface for GitHub Copilot in VS Code. Accessed via Ctrl+Shift+I. Supports custom agents via the agent picker dropdown (Ctrl+Shift+A).

See InfraOps Conductor.

A text-based interface for interacting with software. This project uses several CLIs: Azure CLI (az), Bicep CLI (bicep), Terraform CLI (terraform), GitHub CLI (gh), and PowerShell (pwsh).

A MkDocs Material feature that renders tabbed content blocks using === "Tab Name" syntax. Used in this documentation to show Bicep and Terraform examples side-by-side without duplicating page structure.

🔗 External: MkDocs Material Content Tabs

Runtime context compression technique for agents approaching model context limits. Defines three tiers — full, summarized, and minimal — with per-artifact compression templates. Managed by the context-shredding skill.

📁 See: .github/skills/context-shredding/SKILL.md

A failure-handling pattern in the iac-common skill that defines mandatory stopping rules for deploy agents. Classifies failures into 6 categories with anomaly detection thresholds to prevent cascading deployment failures.

📁 See: .github/skills/iac-common/

A graph where edges have a direction and there are no cycles — you can never follow the arrows back to where you started. In workflow engines, a DAG models task dependencies: each step points to the steps that must come after it, guaranteeing a clear execution order with no infinite loops. This project’s workflow is encoded as a DAG in workflow-graph.json.

Step 3 agent that generates architecture diagrams and Architecture Decision Records (ADRs). Optional step in the workflow. Uses azure-diagrams and azure-adr skills.

📁 Output: agent-output/{project}/03-des-*.{excalidraw,py,png,md}

A Docker-based development environment defined in .devcontainer/. Provides consistent tooling (Azure CLI, Bicep, PowerShell) across all machines.

🔗 External: VS Code Dev Containers

A programming language designed for a specific problem domain rather than general-purpose use. Bicep is a DSL for Azure resource deployment; HCL is a DSL for infrastructure configuration.

A visual diagram showing how data entities relate to each other. Used in the Design step (Step 3) to model data architectures.

The 8-dimension scoring model used to evaluate Ralph Loop runs. Dimensions: artifact completeness, structural compliance, code quality, review thoroughness, WAF coverage, cost accuracy, session state integrity, and timing performance. Composite score 0–100 with letter grades (A–F).

📁 Output: agent-output/{project}/08-benchmark-report.md, 08-benchmark-scores.json

Orchestration agent that drives the Ralph Loop. Executes all InfraOps steps without human gates, with pre-validation, self-correction, challenger reviews, and benchmark collection. Supports both Bicep and Terraform IaC tracks. Invoked via prompt files, not direct @mention.

📁 See: .github/agents/e2e-conductor.agent.md

An experimental conductor variant (01-Conductor (Fast Path)) optimized for simple Azure projects with 3 or fewer resources, single environment, and no custom policies. Combines the Plan and Code steps with a single-pass review for faster delivery.

📁 See: .github/agents/01-conductor-fastpath.agent.md

Azure Policies and organizational rules that affect resource deployment. Discovered during the planning step and documented in 04-governance-constraints.md.

The declarative language used by Terraform to define infrastructure resources. File extension: .tf. Supports variables, modules, data sources, and provider blocks.

🔗 External: HCL Documentation

HIPAA (Health Insurance Portability and Accountability Act)

Section titled “HIPAA (Health Insurance Portability and Accountability Act)”

US regulation governing protected health information (PHI). Azure provides HIPAA-compliant services when properly configured. S04 Service Validation scenario demonstrates HIPAA-compliant architecture.

Azure networking pattern where a central “hub” VNet contains shared services (firewall, VPN gateway) and “spoke” VNets contain workloads. Spokes peer with the hub for connectivity.

The master orchestrator agent that coordinates all steps of the infrastructure workflow with mandatory human approval gates. Implements the Conductor pattern from VS Code 1.109’s agent orchestration features.

📁 See: .github/agents/01-conductor.agent.md

Practice of managing infrastructure through code files (Bicep, Terraform, ARM) rather than manual portal clicks. Enables version control, automation, and repeatability. This project supports two IaC tracks: Bicep (Azure-native DSL) and Terraform (multi-cloud HCL).

A lightweight data interchange format. Used throughout this project for configuration files (agent-registry.json, workflow-graph.json, session-state.json), MCP communication (JSON-RPC), and Azure ARM templates.

🔗 External: JSON Specification

Query language used in Azure Monitor, Log Analytics, and Application Insights. Used for troubleshooting and diagnostics (see S05 Troubleshooting scenario).

🔗 External: KQL Reference

Protocol for extending AI assistants with external tools and data sources. This project uses MCP servers for GitHub (repo operations), Azure (RBAC-aware resource access), Azure Pricing (cost estimation), Terraform Registry (module lookup), and Microsoft Learn (official documentation search and code sample discovery).

📁 See: mcp/azure-pricing-mcp/, MCP Integration

A JavaScript file using modern import/export syntax (as opposed to .cjs which uses require()). Bosun’s codebase uses .mjs files. This project’s validation scripts in scripts/ also use the .mjs extension.

Average time to restore service after an incident. Key SRE metric. Copilot-assisted troubleshooting reduces MTTR by 73-85% (see Time Savings Evidence).

Azure resource that filters network traffic with allow/deny rules. Applied to subnets or NICs. Essential for microsegmentation and defense-in-depth.

PCI-DSS (Payment Card Industry Data Security Standard)

Section titled “PCI-DSS (Payment Card Industry Data Security Standard)”

Security standard for organizations handling credit card data. S04 Service Validation scenario demonstrates PCI-DSS compliant architecture patterns.

Azure feature that assigns a private IP address to a PaaS service (Storage, SQL, Key Vault), removing public internet exposure. Essential for zero-trust architectures.

A document defining the features, functionality, and constraints for a product or project. Ralph uses a prd.json task list to track user stories. In this project, the equivalent is 01-requirements.md.

Azure’s authorization system that assigns permissions based on roles (Owner, Contributor, Reader). Managed through Azure AD / Entra ID. The Azure MCP server is RBAC-aware.

🔗 External: Azure RBAC

An architectural style for web APIs using standard HTTP methods (GET, POST, PUT, DELETE). The governance-discovery-subagent queries Azure Policy assignments via REST API.

A financial metric measuring the gain or loss from an investment relative to its cost. Used in presenter materials to quantify the value of Agentic InfraOps.

A protocol for executing functions on a remote server. MCP servers communicate using JSON-RPC, a lightweight RPC protocol encoded in JSON.

An autonomous, self-correcting E2E evaluation workflow based on the RALPH pattern. Runs all InfraOps pipeline steps without human gates, with built-in self-correction, challenger reviews, and benchmark scoring. Supports both Bicep and Terraform IaC tracks.

📁 See: E2E Testing documentation

A collection of libraries and tools for building applications that interact with a service. Azure SDKs exist for Python, .NET, JavaScript, Go, and Java.

In Azure, a SKU defines the pricing tier and capabilities of a resource (e.g. Standard_LRS for storage, P1v3 for App Service). The Architect agent recommends SKUs based on requirements and pricing data.

A formal commitment from a cloud provider guaranteeing a minimum level of availability (e.g. 99.95% uptime). SLA requirements drive SKU and architecture decisions.

SOC 2 (System and Organization Controls 2)

Section titled “SOC 2 (System and Organization Controls 2)”

An auditing framework for service organisations covering security, availability, processing integrity, confidentiality, and privacy. Azure services hold SOC 2 certifications.

A language for managing relational databases. Azure SQL Database is a managed relational database service used in several example architectures in this project.

An engineering discipline that applies software practices to infrastructure and operations. MTTR is a key SRE metric tracked in this project’s time-savings evidence.

Inventory of all software components in an application, including dependencies and versions. Required for supply chain security. S06 SBOM Generator scenario demonstrates SBOM generation.

Microsoft partner organization that implements Azure solutions for customers. Primary audience for Agentic InfraOps methodology.

A reusable knowledge module stored in .github/skills/ that agents can invoke. Unlike agents, skills don’t have their own chat persona — they provide domain knowledge that agents use. Skills are organized across conventions, document creation, infrastructure patterns, workflow automation, and troubleshooting categories.

📁 See: .github/skills/

A specialized validation agent invoked by other agents for specific tasks (lint, what-if/plan, review). Nine exist: challenger-review-subagent, cost-estimate-subagent, governance-discovery-subagent, bicep-lint-subagent, bicep-review-subagent, bicep-whatif-subagent, terraform-lint-subagent, terraform-review-subagent, terraform-plan-subagent.

📁 See: .github/agents/_subagents/

Key-value pairs applied to Azure resources for organization, cost tracking, and policy enforcement. Baseline tags: Environment, ManagedBy, Project, Owner. Governance constraints may require additional tags. See bicep-code-best-practices.instructions.md or terraform-code-best-practices.instructions.md for the canonical tag rule.

HashiCorp’s open-source Infrastructure as Code tool using HCL (HashiCorp Configuration Language). Supports multi-cloud deployments. In this project, Terraform is the alternative IaC track alongside Bicep, sharing requirements, architecture, and design steps (1-3) before diverging into Terraform-specific planning, code generation, and deployment (steps 4-6).

Provider pin: ~> 4.0 (AzureRM). Backend: Azure Storage Account.

🔗 External: Terraform Documentation

A pluggable Terraform linter that enforces best practices, naming conventions, and resource-specific rules. Used by the terraform-lint-subagent during Step 5 validation.

🔗 External: TFLint

The JSON file that tracks the mapping between Terraform configuration and real-world resources. Stored remotely in an Azure Storage Account for team collaboration. State locking prevents concurrent modifications.

Cryptographic protocol that provides secure communication over networks. This project’s security baseline mandates TLS 1.2 minimum on all Azure services.

The duration a cached value remains valid before being refreshed. The Azure Pricing MCP server uses a 256-entry cache with 5-minute TTL for pricing data and 24-hour TTL for retirement data.

Final testing phase where end users verify the system meets business requirements.

Microsoft’s guidance for building reliable, secure, efficient Azure workloads. Five pillars: Reliability, Security, Cost Optimization, Operational Excellence, Performance Efficiency.

🔗 External: Azure Well-Architected Framework

Azure deployment preview that shows what resources will be created, modified, or deleted without making actual changes. Run with az deployment group create --what-if.

An encrypted network tunnel connecting on-premises networks to Azure virtual networks. Azure VPN Gateway sits in the hub VNet in a hub-spoke topology.

A Windows feature that runs a Linux environment directly on Windows without a virtual machine. Required for Docker Desktop on Windows. The dev container setup guide covers WSL 2 installation.

🔗 External: WSL Documentation

A human-readable data serialisation format used for configuration files. In this project, YAML is used in agent frontmatter (.agent.md), instruction frontmatter (.instructions.md), MkDocs configuration (mkdocs.yml), and GitHub Actions workflows.

🔗 External: YAML Specification

The core Agentic InfraOps workflow: requirementsarchitect → Design Artifacts → Governance → IaC Plan → IaC Code → Deploy → Documentation. Step 3.5 (Governance) runs between Design and IaC Plan. Steps 1–3 and 7 are shared; steps 4–6 diverge into Bicep track (bicep-plannerbicep-codegenbicep-deploy) or Terraform track (terraform-plannerterraform-codegenterraform-deploy). Each step produces artifacts in agent-output/.

📁 See: Workflow Guide

TermFull NameCategory
AADAzure Active Directory (Entra ID)Identity
ADRArchitecture Decision RecordDocumentation
AgentCopilot Custom AgentAI
AKSAzure Kubernetes ServiceCompute
APIApplication Programming InterfaceGeneral
ARMAzure Resource ManagerAzure
AVMAzure Verified ModulesIaC
AVM-TFAzure Verified Modules for TerraformIaC
CAFCloud Adoption FrameworkMethodology
CDNContent Delivery NetworkNetworking
CLICommand-Line InterfaceTooling
DAGDirected Acyclic GraphArchitecture
DSLDomain-Specific LanguageGeneral
ERDEntity-Relationship DiagramDocumentation
HCLHashiCorp Configuration LanguageIaC
IaCInfrastructure as CodeMethodology
JSONJavaScript Object NotationData Format
KQLKusto Query LanguageMonitoring
MCPModel Context ProtocolAI Integration
MJSECMAScript ModuleJavaScript
MTTRMean Time To RecoveryOperations
NSGNetwork Security GroupNetworking
PCI-DSSPayment Card Industry Data Security StandardCompliance
PRDProduct Requirements DocumentDocumentation
RBACRole-Based Access ControlSecurity
RESTRepresentational State TransferArchitecture
ROIReturn on InvestmentBusiness
RPCRemote Procedure CallArchitecture
SBOMSoftware Bill of MaterialsSecurity
SDKSoftware Development KitTooling
SkillCopilot Skill ModuleAI
SKUStock Keeping UnitAzure
SLAService Level AgreementOperations
SOC 2System and Organization Controls 2Compliance
SQLStructured Query LanguageData
SRESite Reliability EngineeringOperations
TFLintTerraform LinterIaC
TLSTransport Layer SecuritySecurity
TTLTime To LiveCaching
UATUser Acceptance TestingQA
VPNVirtual Private NetworkNetworking
WAFWell-Architected FrameworkArchitecture
WSLWindows Subsystem for LinuxTooling
YAMLYAML Ain’t Markup LanguageData Format

See also: FAQ · How It Works · Troubleshooting

Missing a term? Open an issue or add it via PR.