Automating RPC Service Interface Testing: Challenges, Process, and Solutions

In recent years, API automation testing has largely focused on HTTP/HTTPS interfaces because they are close to end‑user interactions and benefit from rich Java and Python libraries, enabling quick framework setup and extensions to functional regression and online monitoring.

With the rise of microservices, RPC frameworks have become popular, but their diverse designs make RPC interface testing more complex. Writing test code for each RPC endpoint is time‑consuming and requires substantial maintenance effort.

The article introduces an approach to automate RPC service interface testing by leveraging the RPC Diff framework, eliminating the need to write separate test code for each RPC method and automatically adapting to interface changes.

A visual overview of the proposed solution is provided, followed by a description of the testing page UI.

The workflow integrates the Beetle platform: when a branch builds successfully, Beetle emits an MQ message that triggers the test service to download code, scan it, compile contracts, and store interface information. Testers then select interfaces on the test page, verify details, modify input templates, and execute tests. The backend generates a unique ID (uniqID), stores request data, returns the ID to the frontend, compiles and runs the test, updates the result, and the frontend polls for completion.

Key technical challenges addressed include converting JSON strings to Java beans, extracting input parameters from logs for both existing and newly changed interfaces, and generating test code using a Velocity‑style template with placeholder substitution.

The article also outlines future extensions aimed at further reducing manual steps and accelerating execution.