All four versions are syntactically and idiomatically fine. Which to use is largely a matter of intended nuance and stylistic choice.
The difference between was/were in the first pair depends on the writer's intended focus...
1a:) was = we saw a single process/action (performed by multiple agents)
2a:) were = we saw multiple agents (performing a single process/action)
For the second pair, I need the linguistic term "patient". The "agent" above is usually the subject of the main verb, whereas "patient" (usually the direct object; terror suspects in this case) refers to who or whatever is affected/acted upon by what the agent does. Much the same nuance / distinction applies to the second pair...
3a:) was = we saw a single process/action (performed on multiple "patients")
4a:) were = we saw multiple "patients" (being subjected to a single process/action)
I should point out that the definite article (the drones) is "optional" in every case (assuming they've already been mentioned earlier in the text; if not, the article definitely shouldn't be included).
Arguably, the use of "the drones" (inclusive of definite article) implies the writer thinks it's more interesting/shocking that they observed drones (not actual human military personnel) engaging in these activities (as opposed to being focused on the fact that these actions were killings, or that the targets were terror suspects). But as I said, it's really a matter of nuance and style.
Just one final observation. OP specifically points out that we don't actually see the drones themselves. But only a pedantic lawyer quizzing a witness would take issue with details like that - it's contextually obvious that the drones were the "active/guilty party" in the event(s) being described, even though they don't feature in the footage being viewed.