Chaos & Fault Injection

Flaky Network & Probabilistic Jitter

Subject your frontend to real-world network instability. Randomly fail requests with configurable failure probabilities (e.g. 25%, 50%, or 75% drop rates) and stochastic jitter delays to verify TanStack Query retries, SWR deduplication, and circuit breakers.

Interactive Probabilistic Chaos Runner

Configure failure likelihood and targeted error codes, then click "Send" repeatedly to observe intermittent failures:

Failure Probability50%

10% (Subtle)50% (Flaky)90% (Outage)

Target Error Status Pool

When chaos triggers, the server randomly picks from these status codes.

Flaky Request (50% Chance of Failure)

Click Send repeatedly. When chaos fires, the server injects 250-1500ms of latency jitter and throws an error.

GET

response.json

{

// Click "Send" above to execute this request against the live server.

}

Chaos Modifiers Reference

Modifier / Header	Type	Valid Values	Behavior
?_chaos=value	Query Param	0.0 to 1.0 (or 0-100)	Probability of random rejection. E.g. `0.3` = 30% failure rate.
X-Simulate-Chaos	HTTP Header	0.5	Header equivalent for CI/CD test runners or automated regression suites.
?_chaos_errors	Query Param	500,502,503	Comma-separated list of target HTTP error codes to randomly select from upon failure.
Network Jitter	Built-in	250ms - 1500ms	Whenever chaos triggers, stochastic latency is automatically added to simulate packet lag.

Resilient Retries with TanStack React Query

Configure automatic exponential retries in your data-fetching hooks to make your UI resilient against transient network drops:

useResilientPosts.ts

// Configuring resilient retries in TanStack Query (React Query)

import { useQuery } from '@tanstack/react-query';

export function useResilientPosts() {

return useQuery({

queryKey: ['posts'],

queryFn: async () => {

// Connect to endpoint with 50% probabilistic chaos injected

const res = await fetch('https://playground.nileslabs.com/api/v1/posts?_chaos=0.5', {

credentials: 'include',

});

if (!res.ok) {

throw new Error(HTTP failure ${res.status}: ${res.statusText});

}

return res.json();

// Retry up to 3 times for transient 5xx / 429 errors

retry: (failureCount, error: any) => {

if (failureCount >= 3) return false;

// Do not retry 400 Bad Request or 404 Not Found

return !error.message.includes('400') && !error.message.includes('404');

// Exponential backoff: 500ms, 1000ms, 2000ms

retryDelay: (attemptIndex) => Math.min(1000 * 2 ** attemptIndex, 10000),

});

}

Want to inspect session telemetry?

Monitor active memory quotas, storage consumption, and visitor session identities.

View Session Quotas

PreviousRate-Limit Simulator

NextSession Quotas & Activity