I had a PLC laying around (Allen-Bradley Micrologix 1500)...but you can get them on ebay for around $100 to $150 used (they just went obsolete so prices may be skyrocketing). The AB programming software is expensive so I wouldn't do that unless I had work access to it (fortunately, in this case, I do). There are other affordable PLCs that come with free programming software. As I said earlier, my intent was only to debounce the limit switch with some logic so it would have to reach a steady state for a few seconds before the PLC would turn off the pump. It migrated from there into a completely overthunk, overdone piece of hardware - but it works great. I later put in some sequencing for the restart of the pump (so it wouldn't have any vacuum on it's input when it started to prolong the health of the pump) and I setup a series of 4 switches that I could use as a timer so I could have the pump quit cycling after some period (after I had gone to bed or work, etc.).

For the vacuum stage sequencing, I installed a smaller reservoir (1.5" PVC about 8 inches long) on the input side of a three-way Mac pneumatic valve (110v). My vacuum pump intake is tee-d into this smaller reservoir. On the other side of the Mac valve is the primary vacuum reservoir that consists of three linked lengths of 6" PVC tubing (and the adjustable vacuum limit switch). The Mac valve switches between venting the smaller reservoir to the atmosphere (through a very fine and restricive brass bead filter) and connecting the small reservoir to the big one. When the pump is turned off and the Mac valve disengaged, the small reservoir takes about 20 or 30 seconds to vent through the restrictive filter on the intake of the mac valve. When the limit switch sets the demand for more vacuum, the PLC fires the pump relay. The pump starts but only initially sees the atmospheric condition in the small reservoir (instead of being hit instantly with the nearly full vacuum in the reservoir). After four seconds, the smaller chamber has reached a pretty significant level of vacuum (the brass filter doesn't let much air flow through) and then the PLC fires the Mac Valve to connect the small reservoir and the primary reservoir (bypassing the brass filter so everything is now sealed to the pump). The pump is now pulling both reservoirs down until the limit switch is satisfied and drops the demand signal to the PLC. On shut-down, the PLC disengages the mac valve to make sure the main reservoir is sealed and then the PLC turns off the pump about one second later. At this point, the smaller reservoir is slowly venting to the atmosphere so it's ready to start over again.

With the timer, I have 4 switches that set a binary timer. When the timer is active, I have an indicator light that flashes once a second. The indicator light also flashes very quickly when the demand switch has changed state to make it easier to adjust the limit switch for different vacuum settings (which I found to be important since the on and off sequence became complicated and is not instantaneously obvious). The switches tie into discrete PLC inputs and any change in state of any of the switches resets the timer to the new value. They're set as follows

0000 - no timer..pump cycling full-time on the limit switch as long as the unit is plugged in and has power.
0001 - 1 hour off delay
0010 - 2 hour off delay
0011 - 3 hour off delay
0100 - 4 hour off delay
0101 - 5 hour off delay
0111 - 6 hour off delay
1000 - and so on.....
....
1111 - the limit switch is ignored and the pump is engaged 100% of the time

And yes, this was absolute and complete over m-f'n kill...but this was back when I had time to play!