Integration tests with test server in GraphQL - Step-by-Step Execution

Concept Flow - Integration tests with test server

Start Test Setup

↓

Launch Test Server

↓

Send GraphQL Query/Mutation

↓

Receive Response

↓

Assert Expected Result

↓

Shutdown Test Server

↓

End

Integration tests start by launching a test server, sending GraphQL requests, checking responses, and then shutting down the server.

Execution Sample

GraphQL

startTestServer()
response = sendGraphQLQuery('{ user { id name } }')
assert response.data.user.id == '1'
stopTestServer()

This code launches a test server, sends a query to get user data, checks the user id, then stops the server.

Execution Table

Step	Action	Request/Query	Response	Assertion	Result
1	Start test server	-	-	-	Server running
2	Send GraphQL query	{ user { id name } }	{ "data": { "user": { "id": "1", "name": "Alice" } } }	-	Response received
3	Assert user id	-	-	response.data.user.id == '1'	Pass
4	Stop test server	-	-	-	Server stopped
5	End test	-	-	-	Test completed successfully

💡 Test ends after server stops and assertions pass

Variable Tracker

Variable	Start	After Step 2	After Step 3	Final
serverStatus	stopped	running	running	stopped
response	null	{ user: { id: '1', name: 'Alice' } }	{ user: { id: '1', name: 'Alice' } }	{ user: { id: '1', name: 'Alice' } }
assertionResult	null	null	true	true

Key Moments - 2 Insights

Why do we need to start and stop the test server in integration tests?

Starting the server (Step 1) creates a real environment to test against, and stopping it (Step 4) cleans up resources. Without this, tests might fail or interfere with each other.

What happens if the assertion fails during the test?

If the assertion (Step 3) fails, the test stops and reports failure. The server should still be stopped to avoid leftover running processes.

Visual Quiz - 3 Questions

Look at the execution table, what is the server status after Step 2?

A starting

B stopped

C running

D unknown

💡 Hint

Check the 'serverStatus' variable in variable_tracker after Step 2

At which step is the GraphQL query sent to the test server?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Look at the 'Action' column in the execution_table for sending queries

If the assertion fails, what should happen next?

A Test stops and server is stopped to clean up

B Test continues without stopping the server

C Server restarts automatically

D Nothing happens, test passes anyway

💡 Hint

Refer to key_moments about assertion failure and server cleanup

Concept Snapshot

Integration tests with test server:
- Start a test server to simulate real environment
- Send GraphQL queries or mutations
- Check responses with assertions
- Stop the server to clean up
- Ensures backend works end-to-end

Full Transcript

Integration tests with a test server involve starting a server that mimics the real backend environment. Then, GraphQL queries or mutations are sent to this server. The responses are checked using assertions to confirm the backend behaves as expected. After tests run, the server is stopped to free resources. This process helps verify that all parts of the system work together correctly.

Practice

1. What is the main purpose of using a test server in GraphQL integration tests?

easy

A. To speed up the production server

B. To run queries and mutations safely without affecting real data

C. To replace the database permanently

D. To generate random data automatically

Solution

1. Step 1: Understand the role of a test server

   A test server is a safe environment that mimics the real server but does not affect actual data.

2. Step 2: Identify the purpose in integration tests

   Integration tests use the test server to check if queries and mutations work together correctly without risk.

3. Final Answer:

   To run queries and mutations safely without affecting real data -> Option B

4. Quick Check:

   Test server = safe testing environment [OK]

Hint: Test server isolates tests from real data changes [OK]

Common Mistakes:

• Thinking test server speeds up production
• Confusing test server with permanent database replacement
• Assuming test server auto-generates data

2. Which of the following is the correct way to start a test server for GraphQL integration tests using Apollo Server?

easy

A. const server = ApolloServer(); server.startServer();

B. const server = ApolloServer(typeDefs, resolvers); server.run();

C. const server = new ApolloServer({ typeDefs, resolvers }); await server.listen();

D. const server = new ApolloServer({ typeDefs, resolvers }); await server.start();

Solution

1. Step 1: Recall Apollo Server setup

   Apollo Server requires creating an instance with typeDefs and resolvers, then calling start() before listen().

2. Step 2: Identify correct method to start server

   The correct method to start the server is await server.start(); before running listen().

3. Final Answer:

   const server = new ApolloServer({ typeDefs, resolvers }); await server.start(); -> Option D

4. Quick Check:

   Use server.start() before listen() [OK]

Hint: Remember to call await server.start() before listen() [OK]

Common Mistakes:

• Calling listen() without starting server
• Using incorrect constructor syntax
• Assuming server.run() or startServer() exist

3. Given this test code snippet for a GraphQL query on a test server:

const result = await server.executeOperation({ query: `query { user(id: 1) { name } }` }); console.log(result.data.user.name);

What will be printed if the user with id 1 has the name "Alice"?

medium

A. undefined

B. null

C. "Alice"

D. Error: user not found

Solution

1. Step 1: Understand executeOperation result

   executeOperation returns an object with data containing the query result if successful.

2. Step 2: Check the query and expected data

   The query requests user with id 1 and its name. If user exists with name "Alice", result.data.user.name will be "Alice".

3. Final Answer:

   "Alice" -> Option C

4. Quick Check:

   Query result matches user name "Alice" [OK]

Hint: executeOperation returns data object with query results [OK]

Common Mistakes:

• Expecting undefined if user exists
• Confusing null with undefined
• Assuming error thrown instead of null result

4. You wrote this test code but it throws an error:

const result = await server.executeOperation({ query: `mutation { addUser(name: "Bob") { id } }` });

What is the most likely cause?

medium

A. The mutation name is incorrect or not defined in schema

B. The query should be a GET request, not mutation

C. executeOperation cannot run mutations

D. Missing await keyword before server.executeOperation

Solution

1. Step 1: Check mutation usage in test

   executeOperation supports mutations, so that is not the issue.

2. Step 2: Verify mutation name and schema

   If mutation name addUser is not defined in the schema, the server throws an error.

3. Final Answer:

   The mutation name is incorrect or not defined in schema -> Option A

4. Quick Check:

   Mutation must exist in schema to run [OK]

Hint: Check mutation name matches schema exactly [OK]

Common Mistakes:

• Thinking executeOperation can't run mutations
• Forgetting to await executeOperation
• Confusing query and mutation types

5. You want to write an integration test that checks if a mutation correctly adds a user and then a query fetches that user. Which sequence correctly tests this on a GraphQL test server?

hard

A. Run mutation with executeOperation, then run query with executeOperation, check query result matches mutation data

B. Run query first, then mutation, then check mutation result

C. Run mutation and query in parallel without waiting, then check results

D. Only run mutation; query testing is unnecessary in integration tests

Solution

1. Step 1: Understand integration test flow

   Integration tests verify that mutations affect data and queries reflect those changes.

2. Step 2: Correct test sequence

   First run mutation to add user, then query to fetch user, then compare results to confirm correctness.

3. Final Answer:

   Run mutation with executeOperation, then run query with executeOperation, check query result matches mutation data -> Option A

4. Quick Check:

   Mutation then query to verify changes [OK]

Hint: Test mutation first, then query to confirm data change [OK]

Common Mistakes:

• Running query before mutation
• Running mutation and query in parallel without order
• Skipping query test after mutation