They didn't know they went too far at the time, and it didn't become obvious for years. But they knew they were getting close at least, and stopped. We'd be worse off if they hadn't denied subsequent projects.
That's just bad engineering. You should know how many pipes and pumps you have of what kind and where, and from that know how many residents it can support. Then you can predict when the infrastructure will become insufficient and start the upgrade in time to finish it before that happens. With enough margin of safety that you're still in the clear even if your estimates are off within the typical range that estimates have been off for similar past projects.
It's no surprise that some places get that wrong from time to time, but the solution is to do better next time. If you're at the point that you have to halt housing construction because your infrastructure can't keep up with it, the point where you screwed up was several years in the past.