Cop limited drops. Every time.

What a drop day costs

Sneaker releases are short, concurrent and latency-sensitive. Fifty tasks is a typical setup for one release:

The split most operators land on: ISP addresses for the tasks that must complete a checkout, because a static address with low latency finishes the form, and residential for breadth when a site rewards having many distinct addresses in the queue rather than a fast one.

One session per task, held through checkout

A checkout that changes address halfway through is a checkout that fails. Give each task its own session id and a lifetime longer than the queue:

Python: a stable address per task

def task_proxy(task_id, country="US", city=None):
    login = f"login_c_{country}"
    if city:
        login += f"_city_{city}"
    login += f"_s_{task_id}_ttl_1h"
    return f"http://{login}:password@proxy.sotaproxy.com:10000"

# 50 tasks, 50 separate addresses, each stable for the whole release
tasks = [task_proxy(i) for i in range(1, 51)]

• Pin the region for regional releases. A release limited to the US west coast is not reachable from an address in Frankfurt, whatever the site says.
• One hour of sticky covers queue plus checkout on most releases. Fifteen minutes does not, and the session will die inside the payment step.
• Buy the addresses before the release, not on the morning of it. Stock in a specific country runs out, and drop days are when everyone discovers that.
• Test the full flow the day before on one task. A configuration that reaches the product page is not the same as one that completes payment.

What loses the checkout