writing-integration-tests

To write integration tests for cloud services, you must declare external service dependencies clearly and enforce the `@pytest.mark.integration` marker for proper test execution and isolation within the SCT framework.

The best way to manage resource cleanup in integration tests is to explicitly terminate all provisioned resources, such as EC2 instances, after execution to prevent cloud resource leaks and avoid unexpected cost overruns.

You mark a pytest test for conditional execution by using the `@pytest.mark.integration` marker and explicitly declaring Docker or Kubernetes as external service dependencies in the test documentation.

Your SCT integration test fails to execute properly if it lacks the required `@pytest.mark.integration` marker or has undeclared external service dependencies, causing the framework to skip or misconfigure the test environment.

Yes, when provisioning AWS resources in integration tests, you need to handle AWS credentials explicitly within the test logic to ensure secure provisioning and successful verification of the cloud instances.

Yes, you can use this guidance to test interactions across GCE and OCI, because it supports declaring dependencies for multiple external services including AWS, Azure, Docker, and Kubernetes environments.