No per-IP concurrency limit on /convert, parallel uploads exhaust Pillow threads (DoS)

[anshul23102]

Bug Description

The /convert endpoint applies no concurrency limit per client. Sending 50 simultaneous PDF conversion requests saturates the Flask thread pool and Pillow worker threads, causing the server to become unresponsive to other users.

Steps to Reproduce

1. Run the following script:

import concurrent.futures, requests
def upload():
    with open('test.pdf','rb') as f:
        return requests.post('http://localhost:5000/convert',
                             files={'file': f}).status_code
with concurrent.futures.ThreadPoolExecutor(50) as ex:
    print(list(ex.map(lambda _: upload(), range(50))))

2. Observe: server CPU reaches 100% and subsequent requests time out.

Root Cause

No request queue, semaphore, or thread-pool size cap in the Flask conversion handler.

Impact

Denial of service: one script can make the service unavailable to all other users.

Proposed Fix

from threading import Semaphore
CONVERT_SEM = Semaphore(4)  # max 4 concurrent conversions

@app.route('/convert', methods=['POST'])
def convert():
    if not CONVERT_SEM.acquire(blocking=False):
        return jsonify({'error': 'Server busy, try again later'}), 503
    try:
        return do_convert()
    finally:
        CONVERT_SEM.release()

[anshul23102]

@Durgeshwar-AI Hi! I came across this issue and I would love to work on it as part of GSSoC '26.

I have reviewed the problem description carefully and am confident I can deliver a well-tested, production-ready fix within the contribution timeline. Please let me know if you need any additional context or have specific implementation preferences before I get started.

Could you please assign this issue to me?

/assign
/claim