The application is currently busy processing 100 requests. It is rejecting further requests.
Maybe you're being DDoS, maybe the application is very slow or blocked, maybe there are too many users and too little hardware resources for the application. Can't tell which case this is without further information.
Look into this. I don't know what WSGI tools you are using to give you the exact settings for your environment.
Increase the number of workers/processes. Each worker can process one request at a time. Multiple workers will be able to process multiple requests in parallel, usually 1 worker per core.
Increase the listening queue to 500. A moderate amount of users could generate bursts of a hundred requests at times. It takes a bigger queue to handle that amount of traffic, assuming there is enough resources to process these requests shortly.
Check your application. Make sure it's working and efficient. There are a hundred requests pending, is this normal? The application might be slow, hanging or lacking resources.