I have a disagreement with my product owner (PO). Our team develops APIs in Mulesoft. We've got a release coming up and PO wants to release one of the APIs that has *some* working endpoints, but other endpoints in that same API have some open bugs.

Given that it's *unlikely* that those broken endpoints will be used, does it seem like good practice to release the API with known bugs in it to hit the deadline?

Understand that we're not just releasing the working endpoints. We're releasing all of it, bugs and all. PO's logic is that those broken endpoints won't be used therefore it's fine to send known bugs into production.

I just need some advice in dealing with this

Maybe don’t ship the buggy endpoints with the next release?

put it in the API documentation.

if nobody ever uses it, nobody ever sees it. and when you deploy a fix, the warning disappears as well.

but when someone actually tries to use it, they at least get a heads-up

Makes no sense to ship broken endpoints. They might even need breaking API changes to overcome the issues. Should not be too hard to leave all the code but not publish the broken endpoints.
If it is hard you have an architectural problem.

If you _have_ to ship all or none of the endpoints then you either have an architectural problem (like @hjk101 said) or your deployment strategy is lacking in something (like a feature flag for the prod, UAT and dev environments for example).

In general, regarding bugs and economical interests, they must be balanced.

Trying to make the best bug free system is useless if a competitor gets to market first and gets the majority share.

I personally would make sure that those aren't security related bugs. Annoying is better than data loss. But not even that is a clear cut. WhatsApp shipped with no and then later with broken encryption and was shortly after sold for 19 billion dollars.

I would have fucked that one up. I wouldn't have shipped it without working encryption, but WhatsApp was more concerned about market share. Hence, I do not have 19 billion dollars and they do.

It's always a gamble. Your instincts are good. But an economical view is important. Listen to the advice of the others, create feature flags, turn off the endpoints and ship.