I experimented with the Compdoc circuits, too, until I realized that his tachometer circuit, designed for a 12 pole alternator, wasn't working so well on a 4-pole motor conversion. The messy waveform basically precludes me from getting RPM info just by tapping the AC leads.
Anyway, back to the topic at hand...
Assuming the inverter is actually drawing 1kW, then maybe the discharge test is telling you that you aren't getting 450 Ah out of the bank, but proving that is a whole different matter. Discharging down to 21V, at the rate specified by the manufacturer for load tests, may be a better way to demonstrate the practical capacity of your bank. Without a data sheet, though, all you have to go by is the generic data that I gave you before. Bear in mind that the capacity of the bank - the "450" number that you have, only applies to a specific rate of discharge, probably the 1/20 rate since they are knock-offs. You are using a 1/12 rate, so naturally it will pull the voltage lower, sooner, and the capacity won't be as great anyway. Your battery bank may be within spec, even though your early test results don't seem to show it at first glance.
One or two deep cycles won't kill the battery bank - provided that you don't do it often, don't go too far down, and that you actually use the resulting knowledge to take better care of the batteries, by avoiding excessively discharging them for the rest of their life cycles. On the other hand, deep-discharges on new batteries aren't necessary unless you have reason to think there's a problem. All the rules and recommendations for good battery care will probalby do you fine.
It's still fun to try, but do you have a "Peukert" chart yet to make sense of the data?