Technique: Condition-Based Waiting

Concept: Never assume "2 seconds is enough." Computers vary in speed. Network latency varies. Hardcoded sleeps are the #1 cause of flaky debugging and CI failures.

The Anti-Pattern

# ❌ BAD: Guessing
submit_form()
time.sleep(2) # "Should be done by now..."
assert_success()

The Pattern: Polling with Timeout

Instead of waiting for time, wait for truth.

Structure

# ✅ GOOD: Knowing
submit_form()

start = time.time()
while time.time() - start < 10:  # Maximum Timeout (Safety Valve)
    if check_if_success_message_is_visible():
        break  # Success! Exit immediately (Fast)
    time.sleep(0.1)  # Tiny polling interval
else:
    raise TimeoutError("Form did not submit within 10 seconds")

assert_success()

Implementation Strategies

1. UI Testing (Selenium/Cypress)

• •Don't use: sleep(5000)
• •Use: await waitFor(() => expect(element).toBeVisible())
• •Why: If it loads in 0.1s, the test finishes in 0.1s. If it takes 4s, the test passes.

2. Async Backend Jobs

• •Don't use: "Wait 5 seconds for the queue to process."
• •Use:
  python

  # Poll the database for the status change
  wait_for_condition(
      lambda: db.get_order(id).status == 'PROCESSED',
      timeout=10
  )
  

3. API Consistency (Eventual Consistency)

• •Scenario: You create a user, but GET /users doesn't show them yet.
• •Fix:
  python

  def get_user_safely(user_id):
      # Retry 3 times with exponential backoff
      for i in range(3):
          try:
              return api.get_user(user_id)
          except NotFound:
              sleep(2 ** i) # 1s, 2s, 4s
      raise NotFound
  

Summary

• •Sleeps are for humans.
• •Conditions are for machines.
• •Always define: "What specifically am I waiting to happen?" and wait for that.