> ## Documentation Index
> Fetch the complete documentation index at: https://docs.usejina.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Context layer

> Understand how Jina connects code, infrastructure, runtime behavior, and issue history to reason about change impact.

# Context layer

The context layer is Jina's connected model of the system under review. It maps how code, infrastructure, runtime behavior, and past issues relate so that a pull request can be evaluated in its real operating context.

## What contributes context

* **Code:** changed files, call paths, contracts, state transitions, and dependencies.
* **Infrastructure:** the services and deployment surfaces involved in running the affected flow.
* **Runtime behavior:** how important flows behave in staging or preview environments.
* **Issue history:** past failures and patterns that can make a new change risky.

## How Jina uses it

When a pull request opens, Jina traces outward from the change. It asks which components depend on the changed behavior, which user or system flows cross those components, and where prior evidence suggests a failure is more likely.

This traced impact is the change's **blast radius**. Jina uses it to prioritize a small set of high-value investigations rather than running an unfocused suite or commenting on every suspicious pattern.

## Why it improves review quality

A diff can look locally correct while breaking a state transition, retry path, integration contract, or downstream consumer. The context layer exposes those connections before Jina decides what to execute.

That combination of system context and runtime evidence helps Jina find deeper bugs while reducing findings that are technically possible but unlikely to matter in the real application.

Continue to [Runtime investigation](/concepts/runtime-investigation).
