"It would be very bad design practice to build in failure modes on the pcb to protect against failure of an external part, because when we bought a new pcb it would also fail.
"
Sure, but that isn't quite what I meant, if it's aimed at me. Some parts are certainly designed to fail safe. If you use a tantalum bead capacitor it will usually fail short, where a foil one will fail open. You wouldn't use a tant if it would mean taking out components on another board when it went.
Problem with relay base could be that someone would plug in one with the same connections, but inferior in some way, perhaps even if superior in others.
You can't always say component A is better in all respects than component B.
As I said, a lot of them will have a long list of relevant parameters, which is why it takes a true engineer to make the choice. He has to consider all the factors which he can work out from the circuit design, are relevant .
Something close would happen in plumbing when deciding what pump should replace some obsolete one. We could all write down 6 or 8 parameters which may or may not be important, with maybe a couple we'd have to find out about.
Peak head, peak flow, head/flow at different points in between, noise, duty, starting characteristic, power factor correction required, materials of construction, temperature rating, connector type, service interval, power requirements, self priming, priming plugs fitted, size, cost, availability and on and on. Then you might find an identical pump, just a bit higher head, so it must be better, right?? ?? Well it depends....!
Only a plumber with an idea of electrics (which may not be enough) can make an informed judgement about a replacement pump.
"If it interfaces with the boiler all is well. " But if you don't know everything about it, you can't judge.
A pump's a bit like an amplifier, but it's a damn site easier to choose a replacement pump.