Skip to main content

Retry Storm Test

This benchmark evaluates how Solvix behaves under repeated failures and ensures it prevents uncontrolled retry storms.

What is a Retry Storm?​

A retry storm occurs when:

  • Multiple requests fail simultaneously
  • Each request retries aggressively
  • Retries multiply exponentially
  • System becomes overloaded

This can:

  • Crash services
  • Overload APIs
  • Cause cascading failures

Why This Test Matters​

In production systems:

  • Failures are inevitable
  • Poor retry strategies amplify issues
  • Systems must fail gracefully

Solvix is designed to:

  • Control retry behavior
  • Prevent retry explosions
  • Stabilize failing systems

Test Objective​

To verify:

  • Retry control under failure
  • Backoff effectiveness
  • System stability during repeated failures
  • No exponential retry explosion

Test Setup​

  • Runtime: Node.js
  • Tool: Tinybench
  • Duration: 2 seconds
  • Scenario: Forced failures
  • Transport: Mock (always fails)

Test Code​

import { Bench } from "tinybench";
import { createClient } from "solvix";

const client = createClient({
  baseURL: "https://example.com",
  retry: {
    retries: 3,
  },
  transport: async () => {
    throw new Error("Simulated failure");
  },
});

const bench = new Bench({ time: 2000 });

bench.add("retry storm test", async () => {
  try {
    await client.get("/test");
  } catch {}
});

await bench.run();

console.log(bench.tasks.map((t) => t.result));

Results​

Latency Mean: ~600 ms
Throughput Mean: ~1.64 ops/sec
Retries per request: Controlled (max 3)

Analysis​

Observations​

  • Each request retries only configured number of times
  • No exponential retry explosion
  • Throughput intentionally reduced (controlled behavior)
  • System remains stable

Internal Mechanisms That Prevent Storms​

Solvix uses:

  • Configurable retry limits
  • Backoff strategies (linear / exponential)
  • Failure detection
  • Controlled retry scheduling

Optional: Custom Backoff Strategy​

const client = createClient({
  retry: {
    retries: 3,
    backoff: (attempt) => attempt * 100,
  },
});

This ensures:

  • Increasing delay between retries
  • Reduced pressure on failing services

Real-World Impact​

Solvix prevents:

  • API overload
  • Cascading failures
  • Infrastructure crashes

This is critical for:

  • Microservices
  • Distributed systems
  • High-traffic applications

Best Practices​

  • Avoid high retry counts
  • Always use backoff
  • Combine with circuit breaker
  • Monitor failure rates

Conclusion​

Solvix successfully:

  • Prevents retry storms
  • Maintains system stability
  • Provides controlled failure handling