prefacing this with saying that I think the whole idea of executing people (by any means) is revolting, doesn't work as a deterrent and can't be undone in the event of a miscarriage of justice...
... a large part of the problem is that the drug companies won't supply the 'best' anaesthetic drugs to be used for that intention, so they end up improvising with other anaesthetic/sedative combos and the people doing the job are not people who use these drugs regularly (eg. anaesthetics/intensive care specialists). Broadly, they're using an anaesthetic, then muscle relaxant, then something to stop the heart quickly. The prison population may also not have completely standard physiology so even if you get IV access the dosing may be tricksier and if you don't understand the pharmacology really well it's actually pretty unbearable to think about how badly you could stuff it up. Securing IV access also doesn't seem to be reliably in the skill set of the people doing it, and there are multiple stories in the media of cannulation taking several hours (obesity and illicit IV drug use in the past may make this much harder for the amateur cannulator).
The doses of sedative they use are massive so unless they're stuffing up dilutions or the cannulas are tissuing then they should 'work' for most people and render them unconscious. It looks like the 'standard' dose of midazolam they use is about 50-100x higher than the max we'd probably use routinely for procedural sedation. That dose on it's own would kill most people because lying flat on your back, as they are, you'd obstruct your airway +/- aspirate +/- develop negative pressure pulmonary oedema, become hypoxic and have a cardiac arrest... but that would look extraordinarily ugly, and takes a long time. The muscle relaxant they use will stop any respiratory effort... again that on it's own would work because it stops you breathing but takes a minute or two (depending on the drug) to work and then the ensuing hypoxia takes time to cause cardiac arrest. If you're only using bolus doses of sedation rather than continuous infusion there is a theoretical risk of the muscle relaxant lasting longer than the sedative and of developing awareness whilst paralysed, but that's unlikely with the super high doses of benzos if it's gone in the right place and hypoxia would be affecting the brain by then. The potassium is what stops the heart quickly (concentrated potassium is incredibly painful to inject awake). I guess they probably use the potassium just to get it over and done with quickly (more of a benefit to the spectators really if the sedation has gone into a vein).
The whole process probably looks grim even when it goes to plan. On the way to achieving anaesthesia, the brain goes through different stages... one of which is where it becomes 'excitable' and triggers writhing/jerking movements (anyone who's ever watched their child have a gas induction will have seen this... kids get really wriggly as they're losing consciousness, but they don't know anything about it). 'Proper' IV anaesthetic induction agents work so quickly that you normally pass through this phase of anaesthesia so quickly you don't usually see this phase of excitement. Drugs like midazolam take longer to work and I wonder, if the dose of drug has gone into a vein, then the twitching/movement people have described while watching executions is the excitable stage of anaesthesia rather than really conscious effort... but it's not really something you want to be taking a chance on given what's going to happen next. Another thing that is occasionally sited as being evidence of distress/awareness is when tears are seen. It's actually not uncommon to see the odd tear even during a completely straightforward, routine anaesthetic, without there being any awareness, however tears are obviously emotive and, given the nature of what they're doing, not something you want to take a chance with. There is technology available that can give an idea of the likelihood of a person being aware, but they don't seem to use it routinely (it's not fool proof and takes expertise to use it properly, and given that they seem happy to crack on without even being able to cannulate properly, I'm not at all surprised they don't have the brains to think about depth of anaesthesia monitoring). Just like with the drug suppliers though, even if the motivation existed to demonstrate the sedative was working, I can't imagine the manufacturers of this equipment would want anything to do with using it for this reason.
Last thing that's perhaps interesting to compare is the processes used in physician assisted dying in countries that allow it. The difference in the quality of death is a) having someone who can get a cannula in properly and b) quick acting, 'proper' drugs at appropriate doses/delivered appropriately. (I don't think that's an argument for making either available for judicial execution...but the absence of both should be an argument for stopping it)