Prepare your application

Konfidence's vector model requires changes to your application. These changes let your services communicate with the other services in their vector. They also let your services receive runtime configuration .

An incoming request flows through Konfidence to your services.

An incoming HTTP request reaches your application through the ingress gateway. In a Kubernetes landscape, the gateway adds X-Vector-ID to the routed request. This header identifies the vector handling the request. Your service must forward it on every outbound HTTP call so that downstream requests stay in the same vector. See Access vector data in your application.

Your service can also use the vector ID to retrieve vector-specific configuration, feature flags, and deployment results. It requests this data from the vector data service through the OpenFeature Remote Evaluation Protocol (OFREP), a standard HTTP API. See Read feature flags in your application.

Konfidence fits microservice applications

Konfidence targets distributed applications. Use it if your application meets these requirements:

• Your application uses a microservice architecture.
• You package each service as its own artifact.

Packaging and integration add overhead for your engineering team. Evaluate carefully whether Konfidence fits your use case.

Next steps

Continue with the guides that support your application workflow:

• Types of artifacts — package your services as Helm or Kustomize artifacts.
• Publish artifacts — push your artifacts to an OCI registry.
• Build vectors — define a VectorTemplate to assemble vectors from your published artifacts.