Overview

Vector data is runtime data that belongs to a specific vector deployment. Applications use it to resolve feature flags, authored configuration values, and deployment results by vector ID.

How vector data reaches the runtime

Konfidence stores vector data in the VectorData custom resource. The landscape orchestrator reads the VectorData resource and passes the data to the target runtime.

The runtime must provide a way for applications to read vector data by vector ID. This keeps runtime data scoped to the vector that is currently handling the request or workload.

Kubernetes runtime

In a Kubernetes runtime, an administrator installs the vector data service in the landscape cluster. The Kubernetes landscape orchestrator pushes vector data to the landscape cluster as ConfigMap resources.

The vector data service reads those ConfigMap resources and provides an OpenFeature-compatible API. Workloads use that API to read vector data by vector ID.

Data types

Vector data contains three data types:

Data type	Purpose
Feature flags	Toggle behavior for features or code paths in the application. Multiple services in the same vector can read the same flag values and behave consistently.
Authored config	Pass arbitrary configuration values to the application.
Deployment results	Pass deployment-generated information into the landscape, such as service endpoints, generated URLs, identities, and allocated resources.