Plumber

A general end-to-end pipeline testing bot that supports various CI/CD pipeline tools.

Why do we need a bot for testing pipelines?

Can't we just run the pipeline with tools like Curl and receive the results through alerts? Why do we need a bot for this?

• The execution of the pipeline and receiving results through alerts may seem sufficient. However, it fails to convey certain contexts. For instance, the results of pipeline execution may not always match the test results. In some cases, it might be desirable for the pipeline to fail as a result of certain tests. Simply relying on pipeline execution alerts won't communicate such nuances.
• Complex pipelines often require interaction with the user. There may be expectations regarding how the pipeline should behave when certain actions are taken by the user. This understanding goes beyond merely running the pipeline. It is necessary to comprehend the entire flow and context of the pipeline to conduct thorough testing.

Concepts

There is Specification for pipeline testing, and the bot performs the tests listed in the specification.

A Test refers to the process of fetching the execution status of a pipeline(or stage) and verifying if it matches the expected state. It must match within the specified time limit.

A TestResult is the result of a test. It contains status information such as whether the test passed or failed, and the reason for failure.

How to use

You can use the bot by creating a specification file and running the bot with the file. The bot will perform the tests listed in the specification. It sends the results to the specified notification channel.

make run-bot-cli spec-file=$(pwd)/plumber-bot/tests/0001_valid-pipeline-spec.yaml

How it works

1. The engine triggers an external pipeline tool to execute the pipeline and initiates pipeline testing. It then periodically queries the pipeline execution results.
2. The engine compares the pipeline execution results with the expected results specified in the testing specification.
3. The engine updates the test status accordingly and repeats this process until the pipeline execution is completed or reaches the specified time limit.
4. The engine generates the test results.
5. The notifier sends the test results to the notification channel specified in the testing specification.