ตัวตัดวงจรและกำแพงกั้น

APIs are the backbone of modern applications, but failures are inevitable. Building resilient APIs means designing them to withstand issues and recover gracefully.

In this lesson, we'll explore two powerful resilience patterns: Circuit Breakers and Bulkheads. These help your systems stay stable even when dependencies struggle.

Imagine a real-world electrical circuit breaker. When there's an overload, it "trips" to prevent damage. In software, a Circuit Breaker pattern does something similar for API calls.

It monitors calls to a service. If too many fail, it "opens" the circuit to that service, stopping further calls for a period. This prevents a failing service from being overwhelmed and allows it time to recover.

A Circuit Breaker typically operates in three main states:

• Closed: Normal operation. Calls to the service go through.
• Open: Too many failures detected. Calls are blocked immediately, returning an error or fallback response without hitting the service.
• Half-Open: After a timeout in the Open state, a few test calls are allowed. If they succeed, the circuit closes; if not, it re-opens.

When your application tries to call a dependent service, the circuit breaker intercepts the call and checks its state:

• If OPEN, it fails fast, returning an error instantly.
• If HALF-OPEN, it allows a single test call to see if the service has recovered.
• If CLOSED, it allows the call and monitors its success or failure.

This "fail-fast" approach is crucial for preventing cascading failures.

Implementing Circuit Breakers provides several key advantages:

• Prevents Cascading Failures: A single failing service won't exhaust resources (like threads) in calling services.
• Faster Failure Detection: Consumers get immediate feedback instead of waiting for slow timeouts.
• Service Recovery: Gives struggling services time to stabilize and recover by reducing incoming load.

Think of a ship with watertight compartments, or bulkheads. If one compartment floods, the others remain dry, preventing the entire ship from sinking.

In software, a Bulkhead pattern isolates resources (like thread pools, connections, or memory) for different services or types of requests. This prevents a failure or slowdown in one component from consuming all shared resources.

Bulkheads work by partitioning resources. Common implementation strategies include:

• Thread Pools: Dedicating separate thread pools for calls to different external services.
• Semaphores: Limiting the number of concurrent calls to a specific downstream service.
• Connection Pools: Isolating database connection pools per microservice or feature.

If one service becomes slow or unresponsive, its dedicated resource pool gets exhausted, but other services' pools are unaffected.

Implementing bulkheads provides strong fault isolation and enhances overall system stability:

• Prevents Resource Starvation: A problematic service won't hog all threads or connections, leaving nothing for healthy services.
• Improved Stability: A failure or slowdown in one area is contained, preventing it from spreading across the entire system.
• Better Diagnostics: Easier to identify which specific component is causing resource issues, as its dedicated pool will show contention.

Circuit breakers and bulkheads are often used together for maximum resilience and robustness.

• A bulkhead isolates a service's resources, preventing its failure from affecting others' capacity.
• A circuit breaker then detects failures within that isolated resource, preventing repeated calls to the struggling service.

This layered approach allows systems to degrade gracefully and recover more quickly from partial outages.

You have an API Gateway that routes requests to multiple backend microservices. One microservice, the 'Recommendation Service', starts experiencing very high latency due to a database issue.

Which pattern would you primarily use to ensure that the slow 'Recommendation Service' doesn't exhaust all available threads in the API Gateway, thus preventing other, healthy microservices from being called?

We've explored two essential patterns for API resilience: Circuit Breakers and Bulkheads.

• Circuit Breakers prevent cascading failures by stopping calls to a failing service, allowing it to recover.
• Bulkheads isolate resources (like thread pools) to contain failures within specific components, preventing resource starvation.

By combining these patterns, you can build highly robust and fault-tolerant API systems that gracefully handle faults and maintain stability.