Refers to #50

Refers to #35

This introduces a new optional configuration parameter called
"LogLevelMonitoring" to adjust the log level of those messages related
to the load monitoring of the service.

Fixes #50

This prevents at least the log file from being cluttered up completely
in "normal" operation. As stated in #50 the level for those messages
should be configurable to allow for easy adjustments without the need to
recompile the service itself.

Fixes #38

Since we're now back on using a central location for storing temporary
transfer files, it does make sense again to check for that location to
exist on service startup (refers to #16).

This reverts commit 10640754.

Refers to #16

The name reflects the functionality much better.

Since this is not a common (central) location any more, it doesn't make
sense to check for it during startup or similar situations.

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Refers to #36

On an integer overflow while being in low-load state the probation
counter should not be reset to zero (indicating the load limit was
violated) but rather to the lowest value representing a valid load
(being _probation + 1).

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Refers to #36

Placed in namespace ATxCommon.Monitoring, using its own timer to update
the values.

Refers to #36

This way an email will be sent to an admin if configured so they have a
chance to check what's going on.

Megabytes have shown to be a bit clumsy in the configuration files, so
simply switch to the more useful Gigabytes unit.

Fixes #39

Refers to #36