Skip to content

Glossary ​

Quick definitions for terms used throughout this reference template. Record your org-specific definitions in your own glossary or ADRs.

TermDefinition
ADRArchitecture Decision Record — documents a significant technical choice, status, and consequences
AIPOAI Platform Owner — Bedrock/Q/MCP config, prompt policies, agent audit
ARBArchitecture Review Board — escalated cross-team design decisions
ABACAttribute-Based Access Control — permissions from attributes (tags, tenant, classification)
Accepted ADRADR status meaning the decision is approved and binding for implementation
Approved specSpecification reviewed and signed off — required before SOP-004
Canary deployRolling out to a small prod slice before full promotion
CI / CDContinuous Integration / Continuous Delivery (or Deployment)
Contract testTest verifying API consumer/provider agreement (e.g. Pact)
DASTDynamic Application Security Testing — runtime security scanning
DATA (steward)Role owning classification glossary and catalog quality
DEVDeveloper — implements against approved spec
EMEngineering Manager — escalation, capacity, non-security exceptions
Ephemeral envShort-lived stack per PR for integration/E2E tests
GuardrailAutomated policy that blocks or warns on unsafe code or usage
Human-in-the-loopRequired human approval at defined gates — AI recommends, humans decide
IAMAWS Identity and Access Management
Identity CenterAWS IAM Identity Center (SSO) for workforce access
IRSAIAM Roles for Service Accounts — EKS pod-level AWS permissions
JITJust-in-time — temporary elevation of privileges
KMSAWS Key Management Service — encryption key management
LF-tagsLake Formation tags for fine-grained data lake access
MCPModel Context Protocol — tool/context integration for AI agents
OIDCOpenID Connect — federation for CI/CD to AWS without long-lived keys
OPAOpen Policy Agent — policy-as-code for IaC and pipelines
OTelOpenTelemetry — vendor-neutral metrics, logs, traces
POProduct Owner — priority and acceptance criteria
RACIResponsible, Accountable, Consulted, Informed
RAGRetrieval-Augmented Generation — LLM + document index
AI toolkitCatalog of AI tools and adjacent artifacts (ADR, specs, MCP)
RBACRole-Based Access Control — permissions assigned to roles
RUMReal User Monitoring — client-side performance and experience
SASTStatic Application Security Testing
SCASoftware Composition Analysis — dependency/CVE scanning
SECSecurity Engineer — threat models, scan policy, exceptions
Secrets ManagerAWS service for storing and rotating secrets
SLOService Level Objective — target reliability metric
SOPStandard Operating Procedure — repeatable team process
Spec-drivenImplementation bound to an approved specification/contract
SRESite Reliability Engineering / platform — deploy, SLOs, incidents
SSMAWS Systems Manager Parameter Store
SSOSingle Sign-On
STRIDEThreat modeling framework (Spoofing, Tampering, …)
T1 / T2 / T3Service tiers — increasing rigor for deploy and on-call (GOVERNANCE)
TechDocs / BackstageDeveloper portal patterns for catalog and docs

See also ​

Reference template only — no production code. Adapt in your own project.