Shared dependencies across routers in FastAPI - Performance & Optimization

Performance: Shared dependencies across routers

MEDIUM IMPACT

This affects the server response time and resource usage by controlling how often dependencies are created and reused across different routes.

Using the same dependency in multiple routers

FastAPI

from fastapi import FastAPI, Depends, APIRouter

app = FastAPI()

async def common_dep():
    # expensive setup
    return 'resource'

app_dependency = Depends(common_dep)

router1 = APIRouter(dependencies=[app_dependency])
@router1.get('/items')
async def items(dep=Depends(common_dep)):
    return {'dep': dep}

router2 = APIRouter(dependencies=[app_dependency])
@router2.get('/users')
async def users(dep=Depends(common_dep)):
    return {'dep': dep}

app.include_router(router1)
app.include_router(router2)

The dependency is declared once at the router level and shared, so it runs only once per request for all routes in that router.

📈 Performance GainReduces redundant calls, lowering CPU usage and improving response time.

Using the same dependency in multiple routers

FastAPI

from fastapi import FastAPI, Depends, APIRouter

app = FastAPI()

async def common_dep():
    # expensive setup
    return 'resource'

router1 = APIRouter()
@router1.get('/items')
async def items(dep=Depends(common_dep)):
    return {'dep': dep}

router2 = APIRouter()
@router2.get('/users')
async def users(dep=Depends(common_dep)):
    return {'dep': dep}

app.include_router(router1)
app.include_router(router2)

The dependency common_dep is called separately for each router and each request, causing repeated expensive setup.

📉 Performance CostTriggers redundant dependency calls per request, increasing response time and CPU usage.

Performance Comparison

Pattern	Dependency Calls	CPU Usage	Response Time	Verdict
Separate dependency per route	Multiple calls per request	High	Slower	[X] Bad
Shared dependency at router level	Single call per request per router	Lower	Faster	[OK] Good

Rendering Pipeline

In FastAPI, dependencies are resolved before route handlers execute. Shared dependencies reduce repeated setup work, improving server processing before response generation.

→Dependency Resolution

→Request Handling

⚠️ BottleneckRepeated dependency calls increase CPU and delay response generation.

Optimization Tips

1Declare shared dependencies at router or app level to avoid repeated calls.

2Avoid declaring expensive dependencies inside each route function.

3Use FastAPI's dependencies parameter on routers to optimize resource usage.

Performance Quiz - 3 Questions

What is the main performance benefit of sharing dependencies across routers in FastAPI?

A Increases the number of dependency calls for better caching

B Makes each route load dependencies independently

C Reduces redundant dependency calls per request

D Removes the need for dependencies entirely

DevTools: Network and Performance panels

How to check: Use the Network panel to measure response times for routes using shared vs separate dependencies. Use Performance panel to profile CPU usage during requests.

What to look for: Look for lower response times and reduced CPU spikes when shared dependencies are used.

Practice

1. What is the main benefit of using shared dependencies across routers in FastAPI?

easy

A. To disable dependency injection in routers

B. To make the API slower by adding extra layers

C. To force all routers to use the same path prefix

D. To avoid repeating the same dependency code in multiple routers

Solution

1. Step 1: Understand what shared dependencies do

   Shared dependencies allow you to write a dependency once and apply it to many routers, avoiding repetition.

2. Step 2: Identify the benefit in code maintenance

   This reduces duplicate code and makes the project easier to maintain and update.

3. Final Answer:

   To avoid repeating the same dependency code in multiple routers -> Option D

4. Quick Check:

   Shared dependencies = avoid repetition [OK]

Hint: Shared dependencies reduce repeated code across routers [OK]

Common Mistakes:

• Thinking shared dependencies slow down the API
• Confusing shared dependencies with path prefixes
• Believing shared dependencies disable injection

2. Which is the correct way to apply a shared dependency to an APIRouter in FastAPI?

easy

A. router = APIRouter(shared_dependencies=[Depends(common_dep)])

B. router = APIRouter(dependencies=[Depends(common_dep)])

C. router = APIRouter(dependency=Depends(common_dep))

D. router = APIRouter(shared=Depends(common_dep))

Solution

1. Step 1: Recall the APIRouter parameter for dependencies

   The correct parameter name is dependencies, which takes a list of Depends objects.

2. Step 2: Match the syntax to the correct option

   router = APIRouter(dependencies=[Depends(common_dep)]) uses dependencies=[Depends(common_dep)], which is the correct syntax.

3. Final Answer:

   router = APIRouter(dependencies=[Depends(common_dep)]) -> Option B

4. Quick Check:

   dependencies param = correct syntax [OK]

Hint: Use 'dependencies' param with a list of Depends [OK]

Common Mistakes:

• Using 'shared_dependencies' instead of 'dependencies'
• Passing Depends directly without a list
• Using wrong parameter names like 'dependency' or 'shared'

3. Given this code snippet, what will be printed when accessing /items/42?

from fastapi import FastAPI, APIRouter, Depends

app = FastAPI()

def common_dep():
    print("Common dependency called")

router = APIRouter(dependencies=[Depends(common_dep)])

@router.get("/items/{item_id}")
async def read_item(item_id: int):
    return {"item_id": item_id}

app.include_router(router)

medium

A. "Common dependency called" is printed once per request

B. An error occurs because common_dep is not awaited

C. Nothing is printed

D. "Common dependency called" is printed twice per request

Solution

1. Step 1: Understand how dependencies run in FastAPI

   Dependencies listed in the router's dependencies run for every matching request automatically.

2. Step 2: Analyze the print statement in common_dep

   Each time the endpoint is called, common_dep runs and prints "Common dependency called" once.

3. Final Answer:

   "Common dependency called" is printed once per request -> Option A

4. Quick Check:

   Router dependencies run once per request [OK]

Hint: Router dependencies run once per matching request [OK]

Common Mistakes:

• Thinking dependencies run multiple times per request
• Expecting no output from print statements
• Confusing async with sync dependencies causing errors

4. Identify the error in this FastAPI router setup with shared dependencies:

from fastapi import APIRouter, Depends

def verify_token():
    pass

router = APIRouter(dependencies=Depends(verify_token))

@router.get("/secure")
async def secure_endpoint():
    return {"message": "Secure"}

medium

A. Depends should be inside a list in dependencies parameter

B. verify_token must return a value

C. APIRouter cannot have dependencies

D. The route path must start with a slash

Solution

1. Step 1: Check the dependencies parameter type

   The dependencies parameter expects a list of Depends objects, not a single Depends instance.

2. Step 2: Identify the correct syntax

   It should be dependencies=[Depends(verify_token)] to be valid.

3. Final Answer:

   Depends should be inside a list in dependencies parameter -> Option A

4. Quick Check:

   dependencies param requires a list [OK]

Hint: Always wrap Depends in a list for dependencies param [OK]

Common Mistakes:

• Passing Depends directly without list
• Assuming dependencies param accepts single Depends
• Ignoring syntax errors from wrong dependencies format

5. You want to apply a shared dependency that extracts a user from a token to two routers, router_a and router_b. Which approach correctly shares the dependency without repeating code?

hard

A. Define the dependency in each route function separately in both routers

B. Include get_current_user as a parameter in FastAPI() app initialization

C. Add the dependency to each router using the dependencies parameter with Depends(get_current_user)

D. Use global variables to store the user and access it in routes

Solution

1. Step 1: Understand how to share dependencies across routers

   Using the dependencies parameter in APIRouter allows applying a dependency to all routes in that router.

2. Step 2: Apply the dependency to both routers

   Adding dependencies=[Depends(get_current_user)] to both router_a and router_b shares the logic without repeating it in each route.

3. Final Answer:

   Add the dependency to each router using the dependencies parameter with Depends(get_current_user) -> Option C

4. Quick Check:

   Shared dependencies via router dependencies param [OK]

Hint: Use dependencies param on routers to share common logic [OK]

Common Mistakes:

• Repeating dependency code in every route
• Trying to add dependencies in FastAPI app instead of routers
• Using global variables instead of dependency injection