This answer is under context that the OP's preprint was submitted primarily by him as a game piece in a competition with another group..
We uploaded the manuscript because we were scared of being scooped. We thought that by uploading the manuscript, it would make it impossible for another group to publish the same results.
This answer uses the context that two or more groups were working on similar measurements in competition, simultaneously and independently; likely there was awareness of each other as the competition for this result; and that the only effect on group B (if they even saw the preprint) is 'lighting a fire' under someone to finish the write up the results to try to either beat a submission date to publication (just in case the OP's preprint was prematurely posted to try to bluff the competition) or to try to show how close both groups were in time and get an 'also ran' credit. Notice that there was no mention of which group thought of these experiments first, or giving credit for the ideas.
Yes, It is ethical for group B to submit a paper for work done independently even knowing that there is a 'preprint' floating around from group A who appear completing their independent results simultaneously. Most of the fixup would occur after one or the other published articles appear.
Giving everyone in this particular situation the benefit of doubt; a solid assumption is that each group expects the other to submit any solid results for publication. (and that both would prefer the published articles to be citation-of-record).
It might be nice to cite a preprint in this case as a placeholder. But, the real issue occurs when the first paper gets 'in print' and the still-waiting group can compare submission dates. Will group A or group B do their duty? (As it turns out from additional comments in the OP's post, group B submitted their work before the OP's preprint appeared.. The OP should add a citation to group B in his in-review paper.).
To expect the preprint to be treated as the priority game changer between the two groups in this very artificial situation is not fair. It was posted as a game-piece in the race to establish priority between neck-and-neck groups doing the same experiment. Either group here could have abused a preprint posting (e.g., putting something out that is not 'ready' or full of minor or major errors to establish claims like a patent troll) and then fixing it up in a later version.
If we are using traditions still mostly in play, when it is a 'race' for priority, what will count to the community in establishing priority is when was the paper submitted to the journal. (If someone tries to publish to early, and their work is not ready, there is chance it will get rejected, giving some checks and balances)
Finally -Journals will care about the submission date, not a preprint posting date. OP's paper is still in review, with group B's paper appearing first. It is usually is not a reason to reject group A's paper. Group B's paper had not appeared when the referees started reviewing. Traditionally, what would happen when the dust settles is that group A and/or group B's paper would add citation or corrigendum.
Pretty much for everyone else, if two papers with similar important results appear about the same time, they both will be cited more or less equally and/or together.
The moral being, (in this case) that if one is worried about establishing priority in a race and trying all possible gambits, the (ready to publish) paper should be submitted for publication immediately after being posted on preprint server. Then it has what protections are currently possible under traditional model and any newly evolving models.
*I likely will get negative comments saying that I imply to 'ignore at will citing or acknowledging unpublished (but reliable) preprints'. (No, I am not). But this case seems to be about independent, simultaneous competition where the preprint was used by group A to either 1) pretend they had already had version ready to be submitted for publication to psych out competitors from submitting their own, or 2) indicate to community that version was just submitted for publication and anyone who thereafter submitted would not have the priority when the dust settled.*
In clarification - in later comments to the original posted questions, here is timeline deduced after my original answer was created.
- Group A (the OP) and group B at conference in their subfield.
- OP or member of group A shows their results in their presentation. From OP's comments, it is pretty clear that any 'published' conference abstracts available before or after conference do not present any details of results (which usually means that presentations standard may be work-in-progress - no one in audience is going to complain or ask for a 'talk' retraction later if results or understanding evolves between talk and a final publication.
- Group B submitted their work for publication without sending a preprint to the OP (or posting a preprint on public server).
- OP submits a preprint a few weeks or month later. They make Group B aware of the preprint.
- OP submits preprint for publication (time delay between preprint and submission unknown)
- Group B's previously submitted paper appears in print.
- OP becomes aware of Group B's paper .
- OP asks question to stack exchange (because no matter what, it is a very disappointing to feel as if you were scooped)
- On bright side, We have important results corroborated within a short time from two independent groups. Others in the field will cite both with with confidence.
1) Numerous edits/rearrangements to articulate reasoning related to original posting which is neglectful of the preprint rights in this particular case focusing on group B actions. (I'd appreciate hints on how to clarify, as the negative comments are focused on preprints being neglected as a general class of published goodness )
2) Added the timeline from OPs additional comments. Basically group B never saw preprint before submitting their own paper because it did not exist then.