Long uploads cause API workers to miss hearbeats

[dkliban]

Version
1.103.1

Describe the bug
In order to support uploads of very large RPMs, we have increased the gunicorn timeouts for the requests. When the request lasts a long time, the API worker misses it's heartbeats and it's AppStatus record ends up getting cleaned up. The next time it tries to write a heartbeat, it fails to do so and restarts.