Seeding Integration Environments

In this lesson, you will learn how to create deterministic state in your integration environments by implementing repeatable data-seeding routines.

When writing integration tests using WebApplicationFactory and Testcontainers, a common hurdle is the "dirty state" problem. If your tests rely on data that changes over time, or if one test modifies the database and leaves it in an unpredictable state for the next, your tests become flaky and unreliable. To combat this, you must treat your test data as code: versioned, automated, and strictly repeatable.

Defining Repeatable Seed Routines

Instead of manually inserting records in every test method, encapsulate your seeding logic into a service or a dedicated initialization class. This routine should follow a "Reset-and-Populate" strategy:

1. Purge: Ensure the database schema is clean (or truncated).
2. Seed: Insert a known, minimal set of entities required to validate the system’s behavior.

By using an IDbSeeder interface, you can inject this logic into your test setup. Here is a concise example of a seed method used within a test suite:

public async Task SeedDefaultData(MyDbContext context)
{
    // Ensure we are working with a deterministic baseline
    await context.Database.EnsureDeletedAsync();
    await context.Database.EnsureCreatedAsync();

    var products = new List<Product>
    {
        new Product { Id = 1, Name = "Alpha", Price = 10.00m },
        new Product { Id = 2, Name = "Beta", Price = 20.00m }
    };

    await context.Products.AddRangeAsync(products);
    await context.SaveChangesAsync();
}

Why Determinism Matters

By programmatically defining your seed data, you ensure that every test run starts from the exact same snapshot. This is critical for integration tests; it allows you to write assertions based on fixed values (e.g., expecting a total count of 2 products) rather than attempting to query the "current" (and potentially polluted) state of the database.

When your seeding routine is idempotent, you gain the confidence to run tests in parallel or in any order, knowing that the environment will always be configured consistently before the first request is sent to your WebApplicationFactory.

Key Takeaway: Effective integration testing relies on automated, deterministic seeding routines that reset the environment to a known state before every execution.