Skip to main content
COMPEL Body of Knowledge
Get Certified Platform

COMPEL Glossary / shadow-traffic

Shadow traffic

A deployment pattern in which a new model or prompt version receives a copy of live traffic and produces outputs that are captured for evaluation but not returned to users.

What this means in practice

Enables realistic pre-release testing without user-visible risk; extended in AI contexts to include side-by-side evaluation against the incumbent model.

Synonyms

shadow deployment , mirror traffic , dark traffic

Related Terms Network

This diagram shows how Shadow traffic connects to other concepts in the COMPEL body of knowledge. Each linked node is a term you can explore in turn.

Related terms for Shadow traffic Network graph showing 3 related glossary terms connected to Shadow traffic. Shadow traffic Evaluation… Continuous… SLI/SLO fo…
3 related concepts connected to Shadow traffic.

The COMPEL Glossary Graph visualizes relationships between framework terminology, showing how concepts interconnect across domains, stages, and pillars. Term nodes cluster by pillar affiliation while cross-references reveal semantic dependencies — for example, how risk appetite connects to control effectiveness, model governance, and assurance requirements. This network representation helps practitioners navigate the framework vocabulary and understand that COMPEL terminology forms a coherent conceptual system rather than isolated definitions.

Accessible list of related terms
  • Evaluation harness — The infrastructure that runs capability, regression, safety, and human-review evaluations on an LLM feature on a defined cadence.
  • Continuous delivery (ML) — Automated, governed promotion of models through lifecycle stages — development, staging, production — with gated checkpoints (evaluation thresholds, bias checks, cost thresholds, human approval where required).
  • SLI/SLO for AI — Service-level indicators and objectives for AI systems — including evaluation score, per-task cost, and goal-achievement rate alongside classical availability/latency.

See also

  • Evaluation harness — The infrastructure that runs capability, regression, safety, and human-review evaluations on an LLM feature on a defined cadence.
  • Continuous delivery (ML) — Automated, governed promotion of models through lifecycle stages — development, staging, production — with gated checkpoints (evaluation thresholds, bias checks, cost thresholds, human approval where required).
  • SLI/SLO for AI — Service-level indicators and objectives for AI systems — including evaluation score, per-task cost, and goal-achievement rate alongside classical availability/latency.