Testing machines too often costs money and wastes resources. Testing machines too infrequently can result in missed faults and catastrophic failures. The key is to find the middle ground and this will depend on the machine, its history and its criticality. Some high speed machines can devolve from perfect health to failure in a matter of minutes. These will require continuous monitoring protection systems. Other machines may run for years without problems. These should be tested quarterly or biannually with test frequency increasing as incipient faults begin to appear. Although a small pump may be inexpensive to repair or replace, if its failure causes a million dollars of product to be ruined, it can be said to be quite critical and worthy of more frequent testing.
The best approach is to spend some time researching the machines, their history and their use. Understanding how the machines fail often gives some indication of how long it will take them to fail after faults begin to appear. A good deal of resources can be saved and better appropriated if records are kept over time and the schedule adjusted as more machinery history and knowledge becomes available.