Gotest

Testcontainers-go plays well with the native go test framework.

The ideal use case is for integration or end to end tests. It helps you to spin up and manage the dependencies life cycle via Docker.

Docker has to be available for this library to work.

1. Install

We use gomod and you can get it installed via:

go get github.com/testcontainers/testcontainers-go

2. Spin up Redis

import (
    "github.com/testcontainers/testcontainers-go"
    "github.com/testcontainers/testcontainers-go/wait"
)

func TestWithRedis(t *testing.T) {
    ctx := context.Background()
    req := testcontainers.ContainerRequest{
        Image:        "redis:latest",
        ExposedPorts: []string{"6379/tcp"},
        WaitingFor:   wait.ForLog("Ready to accept connections"),
    }
    redisC, err := testcontainers.GenericContainer(ctx, testcontainers.GenericContainerRequest{
        ContainerRequest: req,
        Started:          true,
    })
    if err != nil {
        t.Error(err)
    }
    defer redisC.Terminate(ctx)
}

The testcontainers.ContainerRequest describes how the Docker container will look.

• Image is the Docker image the container starts from.
• ExposedPorts lists the ports to be exposed from the container.
• WaitingFor is a field you can use to validate when a container is ready. It is important to get this set because it helps to know when the container is ready to receive any traffic. In this case, we check for the logs we know come from Redis, telling us that it is ready to accept requests.

When you use ExposedPorts you have to imagine yourself using docker run -p <port>. When you do so, dockerd maps the selected <port> from inside the container to a random one available on your host.

In the previous example, we expose 6379 for tcp traffic to the outside. This allows Redis to be reachable from your code that runs outside the container, but it also makes parallelization possible because if you add t.Parallel to your tests, and each of them starts a Redis container each of them will be exposed on a different random port.

testcontainers.GenericContainer creates the container. In this example we are using Started: true. It means that the container function will wait for the container to be up and running. If you set the Start value to false it won't start, leaving to you the decision about when to start it.

All the containers must be removed at some point, otherwise they will run until the host is overloaded. One of the ways we have to clean up is by deferring the terminated function: defer redisC.Terminate(ctx).

Tip

Look at features/garbage_collector to know another way to clean up resources.

3. Make your code to talk with the container

This is just an example, but usually Go applications that rely on Redis are using the redis-go client. This code gets the endpoint from the container we just started, and it configures the client.

endpoint, err := redisC.Endpoint(ctx, "")
if err != nil {
    t.Error(err)
}

client := redis.NewClient(&redis.Options{
    Addr: endpoint,
})

_ = client

We expose only one port, so the Endpoint does not need a second argument set.

Tip

If you expose more than one port you can specify the one you need as a second argument.

In this case it returns: localhost:<mappedportfor-6379>.

3. Run the test

You can run the test via go test ./...