Really why?
All armour is hardened. The hardness of mild steel is about 100 BNH, WW2 era homogenous steel armour steel can range from anywhere between 200 to over 400. This gives it better performance against projectiles, even with the same alloy composition. In American trials, half an inch of average Hadfield Manganese steel had a ballistic limit of 1000 f/s when shot at with .45 ACP ball, half-hard Hadfield steel had a limit of just 613 f/s. 3/4 hard Hadfield Manganese steel performed a little better than half, but still not as good as full: 625 f/s. Similarly, half-hard stainless steel (Rockwell C-33) had a ballistic limit of 646 f/s compared to full-hard (Rockwell C-45) ballistic limit of 658. This is, of course, not a linear relationship. "Full hard" in this case is not as hard as the material will go, but the optimal hardness for the alloy.
Let's step back a bit and talk about penetration mechanics. To simplify it a lot, when a projectile hits an armour plate, the kinetic energy carried by the shell has to go somewhere. If the armour is hard and the shell is soft, the shell will deform and not penetrate. If the armour is soft and the shell is hard, the shell will retain its shape and the armour will deform. If the armour is very thick it can deform without being penetrated, so for thicker armour it's okay to be softer. If the armour is thin, it doesn't have far to go, so it is likely going to be harder. There is also nuance here, since armour that isn't ductile can't bend too far without shattering, whereas softer but more ductile armour can absorb energy by deforming and not shatter.
This is all a very complicated mechanism, so let's go back to the T-34 specifically.A special brand of armour (MZ-2, Mariupol Factory #2) was developed to be used in the T-34 tank. The goal of this alloy was to create armour that was not only thick, but tough. The issue of cracking was explicitly addressed in the Mariupol Factory's report on armour production:
2) Броня из стали "МЗ-2" при твердостях 2,9 – 3,2 по Бринеллю обладает высокой бронестойкостью при обстреле снарядами калибра 37 и 45 мм и достаточной вязкостью, обеспечивающей отсутствие трещин и расколов.
2) At Brinell hardness of 2.9-3.2, armour from MZ-2 steel has high toughness under fire from 37 mm and 45 mm shot and sufficient ductility to avoid cracks or shattering.
Here the 2.9-3.2 means the impression diameter of 10 mm tungsten ball, so 363-444 BHN. This is a lot compared to mild steel, but not a lot compared to practice at the time. For instance, the German standard for 30-50 mm thick armour at the time was 279–324 BHN, so softer but not much softer. In practice, this could vary much more than stated in requirements, for example the StuG captured by the British in North Africa had its 30 and 50 mm plates range from 317 to 585 BHN.
But let us return to the T-34. The issue of armour cracking and spalling was closely monitored by TsNII-48 (the Soviet scientific body concerned with armour metallurgy). A large report was published in the fall of 1942 analyzing damage to 177 T-34 tanks. It covers the nature of damage to armour in some detail. The part that interests us is brittle vs ductile damage. Out of 534 hits examined, 289 or 54.1% resulted in just a dent. 225 or 42% resulted in complete penetration of the armour. In this case the armour did not spall or crack, so it was not excessively brittle.
Brittle penetrations were relatively uncommon. Only 3.9% of hits resulted in brittle penetrations: 2.1% large breaches, 0.6% breaches with cracks, 0.6% cracks, and in 0.6% cases the armour plate shattered. No brittle damage at all was caused by 20 and 37 mm gun calibers, 0.2% of all damage was caused by 42 mm guns (this is likely the 42/28 mm Pak 41), 0.6% by 50 mm guns, 0.6% by 75 mm guns, 0.4% by 88 mm guns, and 0.2% by 105 mm guns. Since the larger caliber guns were rarer, it would be more representative to make a percentage of brittle damage sustained by T-34 tanks for each gun: 0% for 20 and 37 mm, 2% for 42 mm, 1% for 50 mm, 6% for 75 mm, 11% for 88 mm, 7% for 105 mm.
To summarize all this data, the T-34 was designed to withstand 37-45 mm class weapons, similar to other medium tanks of its time. The armour was not excessively hardened by comparison to similar vehicles. Wartime data shows that its armour resisted those weapons well without cracking. The percentage of cracking and shattering increased as the caliber of weapons increased. 75 and 88 mm guns that were not considered by the designers originally deal considerably more damage to the tank that results in cracking and spalling. Note that this should not in any way be considered a property unique to the T-34. American, British, and Soviet trials found that the armour of the Tiger and Panther tanks, while thicker and softer, was vulnerable to shattering and spalling under attack from even lower caliber weapons.
Sources and further reading: