OpenMetadata/openmetadata-integration-tests

OpenMetadata Integration Tests

This module contains SDK-based integration tests that run against a real OpenMetadata server using Testcontainers. Tests execute in parallel and are isolated using TestNamespace.

Quick Start

# Run all tests with MySQL + Elasticsearch (default)
mvn test -pl :openmetadata-integration-tests

# Run with PostgreSQL + OpenSearch
mvn test -pl :openmetadata-integration-tests -Ppostgres-opensearch

# Run a specific test
mvn test -pl :openmetadata-integration-tests -Dtest="TableResourceIT"

Available Profiles

Profile	Database	Search Engine
mysql-elasticsearch (default)	MySQL 8.3.0	Elasticsearch 8.11.4
postgres-opensearch	PostgreSQL 15	OpenSearch 2.19.0
postgres-elasticsearch	PostgreSQL 15	Elasticsearch 8.11.4
mysql-opensearch	MySQL 8.3.0	OpenSearch 2.19.0

Writing a New Integration Test

1. Create the Test Class

Extend BaseEntityIT for entity CRUD tests or create a standalone test class:

package org.openmetadata.it.tests;

import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertNotNull;

import org.junit.jupiter.api.Test;
import org.openmetadata.it.util.TestNamespace;
import org.openmetadata.schema.api.data.CreateTable;
import org.openmetadata.schema.entity.data.Table;

public class MyFeatureIT extends BaseEntityIT<Table, CreateTable> {

  @Override
  protected String getEntityType() {
    return "table";
  }

  @Override
  protected Table createEntity(CreateTable request) {
    return SdkClients.adminClient().tables().create(request);
  }

  @Override
  protected CreateTable createRequest(String name, TestNamespace ns) {
    return new CreateTable()
        .withName(name)
        .withDatabaseSchema(SharedEntities.getSchema().getFullyQualifiedName())
        .withColumns(List.of(new Column().withName("id").withDataType(ColumnDataType.INT)));
  }

  @Test
  void myCustomTest(TestNamespace ns) throws Exception {
    CreateTable request = createRequest(ns.prefix("myTable"), ns);
    Table table = createEntity(request);

    assertNotNull(table.getId());
    assertEquals(request.getName(), table.getName());
  }
}

2. Key Concepts

TestNamespace

Every test method receives a TestNamespace parameter that provides unique prefixes for entity names:

@Test
void myTest(TestNamespace ns) {
  String uniqueName = ns.prefix("myEntity");  // e.g., "abc123_myEntity"
}

This ensures tests don't conflict when running in parallel.

SdkClients

Get pre-configured SDK clients for different users:

OpenMetadataClient adminClient = SdkClients.adminClient();
OpenMetadataClient user1Client = SdkClients.user1Client();
OpenMetadataClient botClient = SdkClients.botClient();

SharedEntities

Access pre-created entities for tests:

DatabaseService service = SharedEntities.getService();
Database database = SharedEntities.getDatabase();
DatabaseSchema schema = SharedEntities.getSchema();
User adminUser = SharedEntities.getAdminUser();

3. BaseEntityIT Features

When extending BaseEntityIT, you get these tests automatically:

Test	Description
post_entityCreate_200	Create entity successfully
get_entity_200_OK	Get entity by ID
get_entityByName_200	Get entity by FQN
get_entityNotFound_404	Get non-existent entity
put_entityCreate_200	Create via PUT
patch_entityAttributes_200	Patch entity attributes
delete_entityAsAdmin_200	Delete entity
get_entityListWithPagination_200	List with pagination
test_sdkCRUDOperations	Full CRUD via SDK
... and 30+ more

4. Controlling Test Behavior

Use flags to customize which inherited tests run:

public class MyEntityIT extends BaseEntityIT<MyEntity, CreateMyEntity> {
  {
    supportsPatch = true;          // Enable PATCH tests
    supportsTags = true;           // Enable tag tests
    supportsOwner = true;          // Enable owner tests
    supportsSearchIndex = true;    // Enable search tests
    supportsDomains = true;        // Enable domain tests
  }
}

5. Best Practices

1. Use TestNamespace.prefix() for all entity names to ensure uniqueness
2. Don't clean up entities - TestNamespace isolation handles this
3. Use specific imports - No wildcard imports (import static ....*)
4. Keep tests independent - Don't rely on order of execution
5. Use Awaitility for async operations - Not Thread.sleep()

Awaitility.await()
    .atMost(Duration.ofSeconds(30))
    .pollInterval(Duration.ofMillis(500))
    .until(() -> someCondition());

6. Avoid single-line comments - Write self-documenting code

Project Structure

openmetadata-integration-tests/
├── src/test/java/org/openmetadata/it/
│   ├── auth/           # JWT token generation
│   ├── env/            # Test infrastructure (TestSuiteBootstrap)
│   ├── factories/      # Entity factory classes
│   ├── tests/          # Integration test classes
│   └── util/           # Utilities (SdkClients, TestNamespace)
└── src/test/resources/
    ├── openmetadata-secure-test.yaml  # Test config
    └── *.der                          # JWT keys

Test Infrastructure

Tests use TestSuiteBootstrap (a JUnit LauncherSessionListener) that:

1. Starts database container (MySQL or PostgreSQL)
2. Starts search container (Elasticsearch or OpenSearch)
3. Starts Fuseki SPARQL container (for RDF tests)
4. Starts the OpenMetadata application
5. Initializes SharedEntities

All containers are started once per test run and shared across all tests.

Running in CI

GitHub workflows run these tests on every PR:

• integration-tests-mysql-elasticsearch.yml - MySQL + Elasticsearch
• integration-tests-postgres-opensearch.yml - PostgreSQL + OpenSearch

Tests require the "safe to test" label on PRs.