ELI5: What is an 'API call limit' and why does it matter for moving my data?

Hey folks! 👋 I've been helping a friend migrate their small e-commerce shop from an old, clunky CRM to a slick new one, and we ran headfirst into the "API call limit" wall. It's a super common gotcha, so I thought I'd break it down with a real example.

Think of an API call limit like a "requests per hour" quota on a library's reference desk. You can only ask the librarian (the old CRM's API) so many questions in a given period. When you're migrating, you're essentially asking for all your customer data, order history, etc., one "question" at a time. If you have 50,000 records and need to fetch them individually, that's 50,000 calls. Many systems have limits like 500 calls per hour. At that rate, your migration would take over 100 hours of non-stop calling!

Here's a snippet of naive Python that would blow through a limit fast:

```python
# This is the problematic way
import requests

all_customers = []
for customer_id in list_of_ten_thousand_ids:
response = requests.get(f'https://old-crm.com/api/customers/{customer_id}')
all_customers.append(response.json())
# This makes 10,000 individual API calls!
```

**Why it matters for your migration:**
* **Downtime:** If you hit the limit, the old system might just start rejecting your requests for hours. Your migration stalls.
* **Data Loss Risk:** If you don't handle rate limiting errors properly, you might miss chunks of data.
* **Cost:** Some platforms charge for overages, or you might need to upgrade to a pricier tier just for the move.

**What you can do:**
1. **Check the docs!** Before signing, find the new *and* old system's API rate limits (look for "rate limiting" or "throttling").
2. **Look for batch endpoints.** A good API lets you fetch many records in one call.
3. **Implement pacing.** Add delays between calls. Here's a better approach:

```python
import requests
import time

all_customers = []
for i, customer_id in enumerate(list_of_ids):
response = requests.get(f'https://old-crm.com/api/customers/{customer_id}')
all_customers.append(response.json())
# Pause after every 10 requests to stay under limits
if i % 10 == 0:
time.sleep(1)
```

Always build in robust error handling and plan for the migration to take much longer than a simple "data size / speed" calculation suggests. Ask the new vendor about their migration tools—do they handle rate limiting gracefully?

Happy coding, and may your data move smoothly! 🚀