HTTP timeout handler should not apply to all API requests

[andrewazores]

Originally the timeout handler was added as a way to prevent indefinite hanging when clients requested non-API paths from the webserver, such as web-client asset paths, which did not exist (ex. from a minimal build image). We have better solutions now that result in a proper 404 response in this scenario anyway. If the timeout handler remains in place then it should at least not apply to actual API requests, which may be reasonably expected in many cases to take longer than the default timeout handler time (ex. uploading recording to archives, uploading recording to jfr-datasource, downloading recording from Cryostat).