Teardown hooks in Flask - Step-by-Step Execution

Concept Flow - Teardown hooks

Request starts

↓

Before request handlers

↓

View function runs

↓

Response generated

↓

Teardown hooks run

↓

Request ends

This flow shows how Flask runs teardown hooks after the response is ready, to clean up resources.

Execution Sample

Flask

from flask import Flask, g
app = Flask(__name__)

@app.teardown_request
def teardown(exception=None):
    print('Teardown hook runs')

This code registers a teardown hook that runs after each request to perform cleanup.

Execution Table

Step	Event	Exception	Action	Output
1	Request starts	None	Start processing request	No output
2	View function runs	None	Process request logic	Response created
3	Response ready	None	Prepare to send response	Response sent
4	Teardown hook runs	None	Run teardown_request function	Prints 'Teardown hook runs'
5	Request ends	None	Cleanup complete	No output

💡 Request processing finished, teardown hooks executed after response sent

Variable Tracker

Variable	Start	After Step 2	After Step 4	Final
exception	None	None	None	None
g	Empty	May hold request data	Cleared or unchanged	Empty or reset

Key Moments - 2 Insights

Why does the teardown hook run even if the view function raises an error?

Because teardown hooks run after the response or error handling, as shown in step 4 of the execution_table, ensuring cleanup always happens.

Can the teardown hook access the exception that caused the error?

Yes, the exception parameter in the teardown function receives the error if any, as shown in the variable 'exception' in variable_tracker.

Visual Quiz - 3 Questions

Look at the execution_table, what is printed during the teardown hook step?

A 'Teardown hook runs'

B 'Request starts'

C 'Response created'

D 'Cleanup complete'

💡 Hint

Check the 'Output' column at step 4 in the execution_table.

At which step does the response get sent to the client?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Look at the 'Event' and 'Action' columns in the execution_table for when the response is sent.

If an exception occurs in the view function, what value does the 'exception' variable have in the teardown hook?

A A string 'error'

B None

C The exception object

D It is undefined

💡 Hint

Refer to the 'exception' variable in variable_tracker and the teardown function parameter.

Concept Snapshot

Flask teardown hooks run after each request finishes.
Use @app.teardown_request decorator.
They receive an optional exception argument.
Useful for cleaning resources like DB connections.
Run even if errors occur in the view.
Executed after response is sent.

Full Transcript

In Flask, teardown hooks are functions that run after each request finishes processing. They are registered using the @app.teardown_request decorator. These hooks run after the response is sent to the client, allowing you to clean up resources such as database connections. The teardown function receives an optional exception argument that contains any error raised during the request. This ensures cleanup happens even if the view function raises an error. The execution flow starts with the request, runs the view, sends the response, then runs teardown hooks before ending the request.

Practice

1. What is the main purpose of a teardown_request function in Flask?

easy

A. To handle incoming HTTP requests

B. To define URL routes

C. To start a new request

D. To run cleanup code after each request finishes

Solution

1. Step 1: Understand the role of teardown hooks

   Teardown hooks run after a request ends to clean up resources like files or database connections.

2. Step 2: Match the purpose with options

   Only To run cleanup code after each request finishes describes running cleanup code after each request finishes, which matches the teardown hook purpose.

3. Final Answer:

   To run cleanup code after each request finishes -> Option D

4. Quick Check:

   Teardown hooks = cleanup after request [OK]

Hint: Teardown hooks always run after request ends [OK]

Common Mistakes:

• Confusing teardown with request handling
• Thinking teardown runs before request
• Assuming teardown defines routes

2. Which of the following is the correct way to register a teardown function in Flask?

easy

A. @app.teardown_request
def cleanup(error=None): pass

B. @app.after_request
def cleanup(): pass

C. @app.route('/teardown')
def cleanup(): pass

D. @app.before_request
def cleanup(): pass

Solution

1. Step 1: Identify the correct decorator for teardown

   The correct decorator to register a teardown function is @app.teardown_request, which optionally accepts an error parameter.

2. Step 2: Check function signature

   The function should accept an optional error=None parameter to detect errors during the request.

3. Final Answer:

   @app.teardown_request\ndef cleanup(error=None): pass -> Option A

4. Quick Check:

   Teardown decorator = @app.teardown_request [OK]

Hint: Teardown functions use @app.teardown_request with error param [OK]

Common Mistakes:

• Using @app.before_request instead
• Missing the optional error parameter
• Confusing with @app.after_request

3. Given this Flask code snippet, what will be printed after a request completes successfully (no error)?

@app.teardown_request
def cleanup(error=None):
    if error:
        print('Error occurred')
    else:
        print('Request finished cleanly')

medium

A. Error occurred

B. Request finished cleanly

C. No output

D. SyntaxError

Solution

1. Step 1: Understand the error parameter in teardown

   The error parameter is None if the request finished without errors.

2. Step 2: Trace the if-else logic

   Since error is None, the else branch runs, printing 'Request finished cleanly'.

3. Final Answer:

   Request finished cleanly -> Option B

4. Quick Check:

   error=None means clean finish [OK]

Hint: error=None means no error, so else branch runs [OK]

Common Mistakes:

• Assuming error is always True
• Expecting no output
• Confusing teardown with before_request

4. What is wrong with this teardown function?

@app.teardown_request
def cleanup():
    print('Cleaning up')

medium

A. Decorator should be @app.before_request

B. Using print inside teardown is not allowed

C. Missing the optional error parameter in the function

D. Function name must be 'teardown' exactly

Solution

1. Step 1: Check teardown function signature

   Teardown functions must accept an optional error=None parameter to receive error info.

2. Step 2: Identify missing parameter issue

   The function cleanup has no parameters, so Flask will raise an error when calling it.

3. Final Answer:

   Missing the optional error parameter in the function -> Option C

4. Quick Check:

   Teardown needs error param or it fails [OK]

Hint: Teardown functions must accept error=None parameter [OK]

Common Mistakes:

• Omitting error parameter
• Using wrong decorator
• Thinking function name is fixed

5. You want to close a database connection after every request, but only if it was opened during that request. Which is the best way to do this using a teardown hook?

hard

A. Use @app.teardown_request with a function that checks if the connection exists and closes it

B. Use @app.before_request to close the connection before the request starts

C. Use @app.route to close the connection on a specific URL

D. Use @app.after_request to close the connection without checking if it exists

Solution

1. Step 1: Understand when to close resources

   Resources like database connections should be closed after the request ends, not before or on specific routes.

2. Step 2: Use teardown_request with conditional check

   The teardown function can check if the connection was opened during the request and close it safely.

3. Final Answer:

   Use @app.teardown_request with a function that checks if the connection exists and closes it -> Option A

4. Quick Check:

   Teardown + check connection before closing = best practice [OK]

Hint: Close resources in teardown only if they exist [OK]

Common Mistakes:

• Closing connection before request starts
• Closing without checking if connection exists
• Trying to close on specific route only