Fault Isolation¶

Fault isolation is the design property that limits the impact of a failure to the component or service where it originates, preventing it from cascading to other parts of the system.

In a tightly coupled system , such as monolith, distributed monolith, a failure in one subsystem can bring down the entire application. Fault isolation ensures that the failure of a payment service, for example, does not prevent users from browsing or adding items to their cart.

Common patterns that implement fault isolation: the circuit breaker, that stops sending requests to a failing service, the bulkhead which limits the resources a single consumer can use, preventing it from exhausting shared pools, and timeouts, that prevents a slow downstream from blocking threads indefinitely.

Microservices architecture promotes fault isolation by deploying each service independently. However, it requires explicit design: synchronous call chains without timeouts or circuit breakers produce the opposite: a single slow service that blocks all callers.

Fault isolation is implemented via timeout configuration on HTTP clients, circuit breaker libraries, and queue-based decoupling of non-critical side-effects.

<?php

// Circuit breaker with a library (e.g., Ganesha or php-circuit-breaker)
$circuitBreaker = Ackintosh\Ganesha\Builder::withRateStrategy()
    ->timeWindow(30)
    ->failureRateThreshold(50)
    ->build();

if ($circuitBreaker->isAvailable('payment-service')) {
    try {
        $result = $paymentClient->charge($order);
        $circuitBreaker->success('payment-service');
    } catch (Exception $e) {
        $circuitBreaker->failure('payment-service');
        // Degrade gracefully — queue for retry, show user a message
        $retryQueue->push(new RetryPayment($order->id));
    }
} else {
    // Circuit is open — skip the call entirely, don't block the user
    $retryQueue->push(new RetryPayment($order->id));
}

?>

Documentation

Fault Isolation¶

Navigation

Related Topics