Optimize Code Performance

You're a DevOps engineer at Checkr. A nightly script sweeps every background check that's still in a pending state and polls the vendor API for a status update, so the ops dashboard stays fresh. The list of pending checks has grown into the tens of thousands, and this script, which used to finish in a couple of minutes, now runs for several hours and is starting to overlap with the next night's run.

import json
import re
import time
import requests

def poll_pending_checks(check_ids, config_path):
    processed = []

    for check_id in check_ids:
        with open(config_path) as f:
            config = json.load(f)

        session = requests.Session()
        pattern = re.compile(r"^(clear|consider|pending)$")

        response = session.get(
            f"{config['api_base']}/checks/{check_id}",
            headers={"Authorization": f"Bearer {config['api_token']}"}
        )
        status = response.json()["status"]

        if pattern.match(status) and check_id not in processed:
            processed.append(check_id)

        time.sleep(2)

    return processed

Explain, in concrete terms, why this takes hours instead of minutes at Checkr's current volume, and how you'd optimize it for a run over tens of thousands of pending checks.