Configurable default wait strategy for DockerComposeEnvironment

[treydavis]

I'm running into a small inconvenience related to using distroless containers with DockerComposeEnvironment. Distroless images don't work with the default port wait strategy. This means I have to use withWaitStrategy to override the strategy for my distroless container. However, I would like to keep specific container-related info out of my tests. For example, I'd like to avoid this:

const environment = await new DockerComposeEnvironment(
    composeFilePath,
    composeFile
).withWaitStrategy(
    'some-specific-container',
    Wait.forLogMessage(/.*some log output specific to the container.*/)
).up();

I'd rather create a healthcheck for some-specific-container in my Docker compose file so that all the container-specific configs are defined in one place, inside the container definition.

This approach has led me to wish for a configurable default wait strategy:

const environment = await new DockerComposeEnvironment(
    composeFilePath,
    composeFile
).withDefaultWaitStrategy(Wait.forHealthCheck());

Or, perhaps, an alternative signature to withWaitStrategy (or a separate method name) that accepts a function mapping container to wait strategy:

const environment = await new DockerComposeEnvironment(
    composeFilePath,
    composeFile
).withWaitStrategy((container: Dockerode.Container) => {
    if (container.Labels... === "something") return Wait.forHttp(...);
    if (container.Labels... === "other") return Wait.forSomething();
    return Wait.forHealthCheck();
})

[cristianrgreco]

Hi @treydavis,

This is the default behaviour. If your containers define their own healthchecks, you simply need to tell Testcontainers to use them:

const container = await new GenericContainer("my-container-with-health-check)
  .withWaitStrategy(Wait.forHealthCheck())

There is an issue related to this, which suggests Testcontainers should automatically switch to use Wait.forHealthCheck() if a container has a health check defined: #687