Spring 启动 docker 应用程序端到端测试与 chrome 驱动程序

Question

I already have Unit and Integration tests with maven but I'm trying to move end-to-end testing with angular(frontend) to maven so I can have every layer of the testing pyramid connected within the same maven build but i'm having trouble doing so.

My spring boot application has some dependencies like postgres and keycloak. I was trying to "mock" does dependencies (in unit and integration tests i use postgres with testcontainers) so i can have an isolated environment without depending on external services are running or not.

I managed to open chrome and do some clicks with selenium and chrome driver but all of this is running on my local database and frontend application which i need to have it up and running. I want something that i can integrate with jenkins later on so i need something to work isolated.

I searched on the internet but i could't find any scenario like this. Anyone could help?

Answer 1

You can have Jenkins flow as

+--------------+          +-----------------+       +-----------+
|              |          | Setup Phase     |       |           |
| Trigger Test | -------->| DB setup and    |------>| Run Tests |
|              |          | Any other setup |       |           |
+--------------+          +-----------------+       +-----------+

Your entire end to end tests would be divided into two phases. In the first phase, you would have to configure all the required services like database, caching layer, test credentials for any third-party services like Keycloak. For example, you need to use a running Postgresql or any other databases as required by the service(s). For other third-party services, you can use their sandbox if available otherwise use a different configuration for each test run (also you can have multiple configurations and pick of them based on the other running tests). Some of the third party services would require you to have public domain/URL in such cases, you need to configure your services behind public URL.

Once you have configured entire systems, then you can start hitting API to gather the response using test scripts. There could be some time, where API might not provide you the exact data you're looking in, in such cases you can add new APIs to the code or use some other technique to gather the data, like DB query, running some manual/automated scripts using FTP, etc.