Definitely 2nd the k-deficiency diagnosis.
This is soilless? Just religiously get 10% runoff or more and it really can't build up in the medium unless you are lying about religiously getting 10% or more runoff. Kinda a mathematical certainty as long as you do. It may rest slightly higher than your feed, but it won't rise above that equilibrium when reached.
This is one of the super powers of soilless/hydro - consistent, predictable level of nutes around the roots. This makes diagnosing easy as pie. Simple formula change fixes anything nutrient-related and it goes into effect almost immediately (in hydro it IS immediate on reservoir change).
No worries about buildup causing lockout... if nutes are interfering with each other or not present at sufficient levels, then it is a matter of the formula and nothing else to consider except pH swings -- which and self-respecting soilless/hydro fertilzer should be ph-balance and ph-buffered to avoid that, too.
Also, if you track PPMs over time, you'll quickly learn various upper /lower thresholds at which you'll see deficiencies/toxicities and effectively avoid it in future, too.
IT's not all simple stuff.. fertilization is a culmination of every choice you made since the start. Feed heavier early on and you don't have to feed as heavy later on -- so it's not 100% apples to apples garden to garden.
I show the per-nutrient ppm formula i use in my diaries. At flip it shifts. A week or two later it shifts again -- The only things i do differently is reduce Nitrogen in 2 steps. The most recent formula does lose a few ticks across the board to get down near ~80ppm N, but at this stage, the plant has been well-fed before this point and even if it causes a slight drawdown in the plant of other stuff, it never amounts to a visible deficiency of concern.
In soilless, fertilizaton should be a no-brainer within a grow or two. You should just expect mostly healthy plants seed to harvest, because it is easy to accomplish if systematic about it and avoid the bro-science influences that inundate us.
more than one way to skin this cat (some ways inevitably superior to others but that isn't fleshed out with research yet), so my numbers may not be what you want -- again it's a culmination of what was provided from the start. Any deviation will require slightly different formula down the road than what i needed in my diaries. But, proof on concept is there. The consistency is going on 7 years, so it's not a terribly small sample.
Check out the pdf screenshot i added to recent week, too. It may profoundly change your perception of how to optimally provide nutes to the roots.