Fallback pattern in Microservices - System Design Guide

Problem Statement

When a microservice call fails due to network issues, timeouts, or downstream service errors, the entire user request can fail, causing poor user experience and cascading failures across the system.

Solution

The fallback pattern provides an alternative response or behavior when a service call fails. Instead of propagating the failure, the system returns a default value, cached data, or a simplified response to keep the system responsive and stable.

Architecture

Client

→Service A

↓

This diagram shows a client calling Service A, which calls Service B. If Service B fails, the fallback handler provides an alternative response to Service A, preventing failure propagation.

Trade-offs

✓ Pros

→

Improves system resilience by preventing failures from cascading.

→

Enhances user experience by providing graceful degradation.

→

Reduces load on failing services by avoiding repeated retries.

✗ Cons

→

Fallback responses may be stale or less accurate than live data.

→

Adds complexity to service logic and testing.

→

Improper fallback can mask real issues, delaying fixes.

Use when downstream services are unreliable or have intermittent failures, especially in systems with high availability requirements and user-facing APIs.

Avoid when data consistency is critical and stale or default responses can cause harm, such as financial transactions or critical control systems.

Real World Examples

Netflix

Netflix uses fallback patterns in its microservices to serve cached or default content when personalized recommendation services are down, ensuring uninterrupted streaming experience.

Amazon

Amazon applies fallback logic in its checkout process to show cached product availability when inventory services are temporarily unreachable, preventing cart failures.

Uber

Uber uses fallback mechanisms to provide estimated fares from cached data when real-time pricing services fail, maintaining user trust during outages.

Code Example

The before code calls the profile service directly and fails if the service is down. The after code wraps the call in a try-except block and returns cached data if the call fails or times out, implementing the fallback pattern.

Microservices

### Before applying fallback pattern

def get_user_profile(user_id):
    # Direct call without fallback
    response = call_profile_service(user_id)
    return response


### After applying fallback pattern

class FallbackException(Exception):
    pass

def get_user_profile(user_id):
    try:
        response = call_profile_service(user_id)
        if response.status_code != 200:
            raise FallbackException()
        return response
    except (TimeoutError, FallbackException):
        # Return cached or default profile as fallback
        return get_cached_profile(user_id)

Alternatives

Retry pattern

Retries the failed request multiple times before failing, instead of immediately providing an alternative response.

Use when: Use when failures are transient and likely to succeed on retry without impacting user experience.

Circuit breaker pattern

Stops calls to a failing service after a threshold, then optionally uses fallback; focuses on preventing overload rather than providing alternative data.

Use when: Use when you want to prevent system overload and quickly detect failing services.

Summary

Fallback pattern prevents system failures from cascading by providing alternative responses when service calls fail.

It improves user experience by enabling graceful degradation during outages or errors.

Fallback should be used carefully when stale or default data is acceptable and avoided when data accuracy is critical.