Can not download large files from minio via Jupyter Server Proxy

Running Jupyter Server 1.9.0 via Docker under OSX. I have an image that starts JupyterLab and uses the Jupyter-Server-Proxy extension to add MinIO.

When I log into the MinIO web interface I can upload large files and download smallish ones (<1-2MB). But if I try to download a largish file (~100MB), I get:

[E 2021-07-06 10:36:40.919 ServerApp] {
      "Host": "",
      "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:89.0) Gecko/20100101 Firefox/89.0",
      "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
      "Accept-Language": "en-US,en;q=0.5",
      "Accept-Encoding": "gzip, deflate",
      "Connection": "keep-alive",
      "Referer": "",
      "Cookie": "username-127-0-0-1-8888=\"2|1:0|10:1625567544|23:username-127-0-0-1-8888|44:NjE4MDZkY2ExNmQ0NGE1ZjljNzJiMjBjNTMxM2ZkNjc=|9c563251fbaace553f943e1b56085bb0a7f393aca4955c87e3f975004468eb27\"; _xsrf=2|a3b32fd7|d207e1bdf72df12803c565b723922e6d|1625153033",
      "Upgrade-Insecure-Requests": "1"
[E 2021-07-06 10:36:40.919 ServerApp] 599 GET /minio/download/my-data/foo.deb?token=eyJhbGciOiJIUzUxMiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3NLZXkiOiJtaW5pb2FkbWluIiwiZXhwIjoxNjI1NTY3ODYwLCJzdWIiOiJtaW5pb2FkbWluIn0.OlQeJhkVXCaYs3VLhDe4A00QqrRIzDsR0Lhl3AADbgOtRHyZsEARoyjixU6IDbzsPN0OAuxlxCfT4HA6Ekecvg ( 4.97ms referer=

Basically 599 which seems to be coming from the Jupyter Proxy service AFAICT. There are no other error logs at least visible in my docker logs.

Is there some kind of limit to download requests? I understand upload requests are threshold by Tornado websocket size. But why am I getting a HTTP 599 connection timeout immediately on download of large files? Where is this limit coming from?

My Jupyter Server Proxy config:

        'command': _get_minio_command,  # Basically, minio server /data
        'port': 9000,
        'timeout': 120,
        'new_browser_tab': True,
        'absolute_url': True,
        'launcher_entry': {
            'title': 'MinIO',
            'icon_path': _get_icon_path()

Note, from the minio console - it doesn’t even look like the download initiaties but that JupyterLabs or the Proxy service (more than likely) is killing it.