Create AI Gateway component for Performance testing in isolation

Description

As part of our initiative to shift left with component-level performance testing, we need to identify few prerequisites before we develop a basic pipeline that demonstrates our Proof of Concept (PoC) in action. This issue is to sort out the various prerequisites that would unblock us from creating a pipeline for running a component level performance test on AI Gateway.

Objectives

1. Create the AI Gateway component in isolation (docker script) 1.
2. Gather the API requests that are being sent to AIGW and also the responses returned by AIGW
3. Setup mocking on AIGW
4. Write down k6 performance tests on one of the api to test the AI Gateway component
5. Document the information needed to setup the component

Tasks

1. Set up the isolated AIGW environment
   • Identify how AI Gateway can be deployed
   • Create scripts to deploy AIGW without a full GitLab instance
   • Implement necessary mocks or stubs for AI models
2. Identify various api requests and responses being sent to AIGW
   • Identify the body, header and query parameters required to be passed with the api Requests
3. Setup mocking for AIGW
   • Use a mock server to mock the responses if not already available out of the box
4. Run a basic api test manually to sort out any authentication/authorization issues and get a 200 respose
5. Write a k6 tests for one of the api
6. Run the k6 test locally on a locally installed AIGW to verify that it is working