The gauge controller should be zeroed out.  In the case of atmospheric pressure this generally can be done with just the controller itself.  There sometimes is an adjustment screw or pot that can be adjusted.  Review of the manual accompanying your gauge should provide the location for this.  In the case of the low pressure or high vacuum reading the gauge should be tested or "set".  This can be accomplished by using a reference gauge tube which has a known vacuum pressure in it, or using a vacuum system capable of generating pressures better than 10^-4 Torr.  In the second case the gauge can then be set at it's maximum vacuum pressure number with some degree of confidence.  In the first the gauge is set to the rated reference tube pressure.  These reference tubes should be submitted for recalibration occasionally because they have a tendency to loose pressure over time.

These systems change over time.  If the tube resides in a high vacuum situation it can become contaminated by the gases evolved in the system.  Many times it is oil vapor, but can be just about anything.  There are two "noble" wires in the transducer head that measure the potential drop from one wire to the other.  The two wires present that measure the potential become coated with impurities or get dirty and when dirty the tend to read high toward atmosphere.  Supposedly, it is possible to "clean" these tubes with carburetor cleaner or acetone although our efforts to date have consistently met with failure.  Usually the only thing that can be accomplished is to replace the tube.  The best method is to use the transducer for your pressure measurement and when not in use, to isolate the transducer through use of a valve or stopcock.   We also have gauge guards which is a small filter between the gauge tube and the system.  Some users swear by them.

Other problems that can be present to varying degrees.   A common one is in neon tube production to "fry" the gauge tube when bombarding.  If high voltage is to be supplied into the system being evacuated then the use of a battery operated gauge will eliminate this problem area. The gauge then does not offer a path to ground.  Another solution is to again isolate the gauge tube during the process with a valve of some kind.