Debugging Prompt Templates

**Root Cause:** The component renders before the API call completes. On first render, `users` is undefined, and you're calling `.map()` on undefined.

**The Issue with &&:** Your null check `users && users.map()` doesn't work because `users` is initialized as `undefined`, not `null`.

**Fix:**
```javascript
// Option 1: Initialize as empty array
const [users, setUsers] = useState([]);

// Option 2: Optional chaining with fallback
{(users ?? []).map(user =>...)}

// Option 3: Early return while loading
if (!users) return <LoadingSpinner />;
```

Option 1 is cleanest, initializing as `[]` means `.map()` always works (returns empty array while loading).

**Walking through your code:**

Line 3: `rate = rate`, You're not converting the percentage. When you pass `10`, you're using 10 (1000%) instead of 0.10 (10%).

**The bug:** Your formula treats the rate as a decimal but your input is a percentage.

**Fix:**
```python
def compound_interest(principal, rate, years):
 rate = rate / 100 # Convert percentage to decimal
 return principal * (1 + rate) ** years
```

**Verification:**
- Input: 1000, 10, 1
- Calculation: 1000 × (1 + 0.10)¹ = 1000 × 1.10 = 1100 ✓

**Most Likely Causes (in order of probability):**

1. **Session affinity issue with load balancer**
 With multiple servers and Redis, if sticky sessions aren't configured, requests might hit a server that doesn't have the session cached locally before Redis sync.
 - *Check:* Are you using `connect-redis` with proper configuration?
 - *Fix:* Ensure all session reads go through Redis, not local memory.

2. **Race condition with multiple tabs**
 Multiple tabs refreshing tokens simultaneously could invalidate each other's sessions.
 - *Check:* Log session token changes with timestamps
 - *Fix:* Implement token refresh locking

3. **Redis connection drops under load**
 Intermittent Redis connectivity could cause session lookups to fail.
 - *Check:* Monitor Redis connection errors
 - *Fix:* Add retry logic to session middleware

**Debugging steps:** Add logging for session operations with request IDs to trace the exact sequence when logouts occur.

**Bottleneck Identified:** The JOIN across 3 large tables is doing a full table scan.

**Analysis of your query:**
```sql
-- Your query scans all 2M rows before filtering
SELECT * FROM orders o
JOIN customers c ON o.customer_id = c.id
JOIN products p ON o.product_id = p.id
WHERE o.created_at > '2024-01-01'
```

**Issues:**
1. No index on `orders.created_at`
2. `SELECT *` fetches columns you likely don't need
3. No index on join columns

**Fixes (in order of impact):**

1. Add composite index:
```sql
CREATE INDEX idx_orders_date_customer_product 
ON orders(created_at, customer_id, product_id);
```

2. Select only needed columns:
```sql
SELECT o.id, o.total, c.name, p.title FROM...
```

**Expected improvement:** 8s → 50-100ms with proper indexing