Why Persisted queries in GraphQL? - Purpose & Use Cases

The Big Idea

What if your app could send tiny messages instead of big query texts every time?

The Scenario

Imagine you have a web app where users send many GraphQL queries. Each time, the app sends the full query text over the internet.

This means slow loading and lots of repeated data sent again and again.

The Problem

Sending full queries every time wastes bandwidth and slows down responses.

It also risks typos or changes causing errors, and makes caching difficult.

The Solution

Persisted queries let you save query texts on the server once.

Clients then send only a small ID to run the saved query.

This speeds up communication, reduces errors, and improves caching.

Before vs After

✗ Before

client sends full query text every time

✓ After

client sends query ID only, server runs saved query

What It Enables

Persisted queries make apps faster and more reliable by sending less data and avoiding repeated query parsing.

Real Life Example

A mobile app uses persisted queries to reduce data use and speed up loading, especially on slow networks.

Key Takeaways

Sending full queries each time is slow and error-prone.

Persisted queries store queries once and reuse them by ID.

This improves speed, reduces data use, and lowers errors.

Practice

1. What is the main benefit of using persisted queries in GraphQL?

easy

A. The server stores user credentials for faster login.

B. Clients send only a unique ID instead of the full query, saving bandwidth.

C. Queries are automatically optimized by the client.

D. It allows clients to write queries without validation.

Solution

1. Step 1: Understand what persisted queries do

   Persisted queries store the full GraphQL query on the server with a unique ID.

2. Step 2: Identify the client-server interaction

   Clients send only the ID to run the query, reducing the data sent over the network.

3. Final Answer:

   Clients send only a unique ID instead of the full query, saving bandwidth. -> Option B

4. Quick Check:

   Persisted queries reduce data sent = Clients send only a unique ID instead of the full query, saving bandwidth. [OK]

Hint: Persisted queries send IDs, not full queries [OK]

Common Mistakes:

• Thinking clients send full queries every time
• Confusing persisted queries with client-side caching
• Believing persisted queries store user data

2. Which of the following is the correct way to send a persisted query request in GraphQL?

easy

A. { "id": "12345", "variables": { "userId": "1" } }

B. { "query": "{ user(id: 1) { name } }" }

C. { "mutation": "updateUser" }

D. { "headers": { "Authorization": "token" } }

Solution

1. Step 1: Identify the persisted query request format

   Persisted queries send the unique query ID and variables, not the full query string.

2. Step 2: Match the correct JSON structure

   { "id": "12345", "variables": { "userId": "1" } } sends an ID and variables, which is the correct persisted query format.

3. Final Answer:

   { "id": "12345", "variables": { "userId": "1" } } -> Option A

4. Quick Check:

   Persisted query request = ID + variables [OK]

Hint: Persisted queries use ID field, not full query [OK]

Common Mistakes:

• Sending full query instead of ID
• Using mutation key instead of id
• Confusing headers with query payload

3. Given the following persisted query setup, what will the server return when the client sends { "id": "abc123", "variables": { "limit": 5 } } if the query with ID abc123 fetches the first limit users?

Assume the database has 10 users named User1 to User10.

Options:

medium

A. []

B. [{"name": "User6"}, {"name": "User7"}, {"name": "User8"}, {"name": "User9"}, {"name": "User10"}]

C. [{"name": "User1"}, {"name": "User2"}, {"name": "User3"}, {"name": "User4"}, {"name": "User5"}]

D. Error: Query ID not found

Solution

1. Step 1: Understand the query and variables

   The query with ID 'abc123' fetches users limited by the 'limit' variable, which is 5 here.

2. Step 2: Determine the expected result

   Since the database has users User1 to User10, fetching the first 5 returns User1 to User5.

3. Final Answer:

   [{"name": "User1"}, {"name": "User2"}, {"name": "User3"}, {"name": "User4"}, {"name": "User5"}] -> Option C

4. Quick Check:

   Limit 5 users returns first 5 users [OK]

Hint: Variables control query output; check their values [OK]

Common Mistakes:

• Assuming query returns last users
• Thinking ID is invalid
• Ignoring variables in query

4. A developer tries to use persisted queries but gets an error: Query ID not found. What is the most likely cause?

medium

A. The client sent a query ID that the server does not recognize.

B. The client sent the full query instead of the ID.

C. The server does not support GraphQL.

D. The client forgot to include variables.

Solution

1. Step 1: Analyze the error message

   The error 'Query ID not found' means the server cannot find the query matching the sent ID.

2. Step 2: Identify the cause

   This usually happens if the client sends an ID that was never registered or stored on the server.

3. Final Answer:

   The client sent a query ID that the server does not recognize. -> Option A

4. Quick Check:

   Unknown query ID causes 'Query ID not found' error [OK]

Hint: Check if query ID is registered on server [OK]

Common Mistakes:

• Assuming full query sent causes this error
• Blaming missing variables for this error
• Thinking server lacks GraphQL support

5. You want to secure your GraphQL API by allowing only persisted queries. Which approach best achieves this?

hard

A. Disable persisted queries and use API keys instead.

B. Allow all queries but log those without IDs for review.

C. Require clients to send full queries and IDs together.

D. Reject any request that does not include a valid persisted query ID.

Solution

1. Step 1: Understand API security with persisted queries

   Allowing only persisted queries means the server accepts requests only if they have a valid stored query ID.

2. Step 2: Identify the best enforcement method

   Rejecting requests without valid IDs ensures only approved queries run, improving security.

3. Final Answer:

   Reject any request that does not include a valid persisted query ID. -> Option D

4. Quick Check:

   Only accept valid persisted query IDs to secure API [OK]

Hint: Block requests missing valid persisted query IDs [OK]

Common Mistakes:

• Allowing all queries weakens security
• Sending full queries defeats persisted query purpose
• Relying only on API keys without query control