“Can we quickly upgrade the API while we’re at it?”

Yes, dear client, yes, we could. If we had the budget for it, and if you had the time to test it, and if you had the energy to start a very long email chain between us and a vendor that replies once every sometimes, because the little documentation you do have is outdated by several months – and for a beta version.

I didn’t push back hard enough when this question was thrown at me a couple of months back. They were certain everything was alright, and I trusted them. That is an engineering failure on my part. API v1 was already doing its job perfectly fine, and was confirmed as supported for the time being. But the client wanted to upgrade, because why not, and their one engineering guy radiated the confidence to convince me. Once we put it live on acc, we found out that a different, second, hidden route we depended on had been removed in v2. What a waste of time!

As an engineer, you should recognise scope creep. They pay for one thing; you’re not gonna do a bunch of other small things alongside it. Talk to your manager about it. Ask them for the evidence you need and for their time budget. Only then can you make the commitment.