timeout everything asynchronous

A reasonable amount of software failures (reported by means of log notifications) is a necessary assumption. Therefore, each and every asynchronous operation must have a built-in timeout to be operational on a production server.
Default limits must be quite generous and the configuration options will be implemented per need.

p.s. Low priority until evidence suggests otherwise.
