🚨 Mocking Network Failures, Retries & Edge Cases in Jest

💡 Why Network Failures Feel Personal

When your code fails, you blame yourself.

When the network fails, you blame life.

Everyone has seen:

• ⚠️ Slow responses
• ❌ Random 500 errors
• 🔁 Infinite loading spinners
• 🧨 “Works on my machine” moments
• 🧟‍♂️ API suddenly returning HTML instead of JSON
• 🧊 Freeze → retry → freeze → retry

All of this is deeply relatable.

This is why mocking failures keeps readers hooked — they’ve lived this suffering.

🧪 Step 1 — The Expense Tracker’s “Retry Logic”

Let’s give our app a new ability:

Try again when the API fails.

📁 fetchRetry.js

const axios = require('axios');

exports.fetchExpensesWithRetry = async (url, attempts = 3) => {
  for (let i = 0; i < attempts; i++) {
    try {
      const res = await axios.get(url);
      return res.data;
    } catch (err) {
      if (i === attempts - 1) throw err;
    }
  }
};

If the API works → great.

If it fails → try again.

If it always fails → cry.

⚙️ Step 2 — Mocking Failure → Success

📁 fetchRetry.test.js

jest.mock('axios');
const axios = require('axios');
const { fetchExpensesWithRetry } = require('./fetchRetry');

test('should retry once and succeed', async () => {
  axios.get
    .mockRejectedValueOnce(new Error('Network Error'))
    .mockResolvedValueOnce({ data: [{ id: 1, amount: 500 }] });

  const result = await fetchExpensesWithRetry('/expenses');

  expect(axios.get).toHaveBeenCalledTimes(2);
  expect(result.length).toBe(1);
});

We fake suffering…

so our users don’t have to.

⚠️ Step 3 — Mocking Permanent Failure

Some APIs never want to work.

Like gym memberships — full of hope, zero results.

test('should fail after max retries', async () => {
  axios.get.mockRejectedValue(new Error('Server Down'));

  await expect(fetchExpensesWithRetry('/expenses'))
    .rejects.toThrow('Server Down');

  expect(axios.get).toHaveBeenCalledTimes(3);
});

Now your tests don’t just celebrate success.

They respect failure.

🐌 Step 4 — Mocking Slow Responses (Timers + Promises)

Sometimes your API responds so slowly

that you begin reevaluating your career choices.

We simulate that:

jest.useFakeTimers();

axios.get.mockImplementation(() =>
  new Promise(resolve => {
    setTimeout(() => resolve({ data: [] }), 3000);
  })
);

jest.runAllTimers();

Three seconds gone in one millisecond.

If only life worked like Jest.

🧩 Step 5 — Mocking Weird/Broken Responses

Let’s simulate the things that cause dev trauma:

Empty data

axios.get.mockResolvedValue({ data: [] });

Null data

axios.get.mockResolvedValue({ data: null });

Corrupted JSON

axios.get.mockResolvedValue({ data: "ThisIsNotJSONBro" });

Your tests should see all of this.

Because your users definitely will.

✅ SUMMARY

Today our Expense Tracker became a warrior.

It learned to survive unpredictable network chaos.

We mastered:
 🔥 Mocking API failures
 🔁 Mocking retries
 🐢 Mocking slow responses
 🧨 Mocking corrupted responses
 🚫 Mocking permanent failure

Your tests no longer trust the internet.

They trust YOU.

Next post → Mocking Permissions, Sensors & Device Behaviors
(This one is MASSIVE for React Native devs.)