CRJ200 Captain reported a repeat incident of cabin pressure failure caused an immediate descent and diversion to make a precautionary landing.

Narrative

The initial taxi; takeoff and climb portions of flight were nominal with no incident. When we leveled off in cruise at FL330 our pressurization system began erratically changing the cabin pressurization. The rate of change is displayed on the MFD (Multi-function flight display) and it would fluctuate rapidly in both directions. We saw pressurization rates of change as high as 1200 feet per minute that caused physical discomfort to my ears and one can assume every other person onboard the aircraft. With the cabin pressure fluctuations I noticed that the overall trend of the cabin pressure showed the cabin altitude slowly climbing. Since the pressurization seemed normal throughout the climb we requested a descent to FL240; as I began to look through the QRH (Quick Reference Handbook) for a procedure associated with rapid changes to cabin pressure; but could not find one. We initially started a shallow descent in order to leave some thrust in; and keep bleed air going to the packs. The pressure changes were still erratic and uncomfortable throughout the descent; and the cabin altitude kept slowly increasing although the rate at which it increased seemed to be increasing as well. At this point we were still descending to FL240; and watching our cabin altitude increase and requested a descent to 10;000 feet. Shortly after that request; a 'cabin alt' caution message presented on the EICAS and I began the associated QRH procedure; as we increased our rate of descent while making sure to leave enough thrust in to run the PACKs. The QRH only had us ensure that all of the pneumatic components that are associated with the pressurization system are in the proper position and receiving enough pressure from the engine bleeds. At this point the QRH had me swap pressure controllers; which seemed to help some; but we were also close to 10;000 feet by the time this happened. In the descent the situation seemed to stabilize some; and we were able to verify that even though the 'cabin alt' caution was still on; the cabin altitude was decreasing. Had we not begun a precautionary descent immediately after noticing the issue; I believe the 'cabin alt' caution would have progressed into a warning and we would have dropped the masks. At this point in time we discussed that our threat was now a time threat and began to evaluate options for the rest of the flight. We were dispatched to ZZZ with an alternate; and while we could make it to ZZZ at 10;000 feet; we would land with only 1;200 pounds of fuel on board. I did not like that option as we would barely have fuel for a go around; and would not have enough fuel to divert if the weather changed. We looked at our position; and determined that ZZZ1 had Company service; and ZZZ2 was even closer and a major airport. I sent an ACARS message to MX (maintenance) during the situation in cruise and was told only that they were notified and to document the irregularity in the eAML (Electronic Aircraft Maintenance Logbook) and call MX Control on the ground. I also wanted to expand my team and use all available resources so I sent a message to Dispatch saying we would more than likely need to divert and asked for potential diversion airports. Our Dispatcher recommended ZZZ2 as a potential diversion airport because the weather was good. We later messaged the Dispatcher to send the NOTAMS for ZZZ2; and began to plan a diversion there. In the descent we noticed that the secondary CPCP (Cockpit Cabin/Pressure Controller) was lagging behind the descent rate of the aircraft and the cabin pressure seemed to be higher than it normally would be at 10;000 feet. Because of this we requested a descent to 5;000 feet and a 5 minute delay vector to let the pressurization catch up before we began the approach to ZZZ2. Upon reaching 5;000 feet; the pressurization looked normal and we decided to begin the approach to ZZZ2. I watched the pressurization throughout the approach and realized that while we were slightly pressurized;we were only pressurized to about 200 ft. above field elevation; which is about 400 ft. higher than on a typical flight. We taxied to the gate and parked at the gate without issue. I elected to delay starting the APU until after landing as I did not want to change any variables to the ECS (Environmental Control System) when it was clearly not working properly. After taxiing to the gate I documented the irregularity to the best of my ability in the eAML; placed the [form] on the thrust levers and made a phone call to MX Control.The same airplane that this event occurred in had an almost identical pressurization event to the one we had the day before we flew it. On the previous leg the airplane had returned to field; and spent the night in maintenance. If I am not mistaken; a similar pressurization event had occurred on the previous two flights; almost identically; although I have lost access to the discrepancy log and can not recall the exact number. The plane had spent the night in ZZZ3 maintenance to have the issue fixed; and we were the first flight following the swapping of the second of two CPCPs. The plane was signed off; and Dispatched; but the maintenance issue clearly had not been fixed. The airplane was never test flown. My crew was set up for failure; there was a zero percent chance of that airplane making it safely to ZZZ. Had I known about the issue and the severity I would have refused the airplane. I wrongly assumed that seeing a closed write up; a corrective action and a return to service meant that the airplane had been fixed and was safe to return to service when it clearly was not. The safety management systems of the company failed my crew; failed the passengers; and failed the operation since we had to be rescued from ZZZ2. That airplane should never have had paying passengers/customers put on it before ensuring that it was safe and fixed correctly. The cause of this event was just trying a solution that was thought might work; not appropriately testing the airplane; signing the airplane off; and then flying a revenue flight with paying passengers aboard and just hoping that the problem would go away with the maintenance action that was performed.