Apparently one @Dusk_Foundation machine can wear two hats… and I honestly thought that sounded efficient at first 😭

I was looking through the archive-node setup and got stuck on one thing. Dusk recommends keeping production API infrastructure separate from provisioner duties, even though the same node can technically do both.

I kept thinking, why? If it can handle both jobs, what's the problem?

Then I looked at what those two jobs actually need.

Archive node: historical indexes and API queries. Provisioner: stay synced and participate in consensus.

Okay. So far, still sounds like one machine should be fine.

Then the annoying part showed up.

Query load and maintenance can compete with the resources the provisioner needs for consensus.

So it isn't really “can the machine run both?”

It's more like… what happens when one side gets busy?

And that changed my thinking a bit. I was counting the hardware twice because it could technically do both jobs.

But the resources aren't infinite just because the machine has two roles.

Maybe the separation isn't really about needing another machine.

Maybe it's about not letting API work become the thing that gets in the provisioner's way.

@Dusk_Foundation $DUSK #DUSK #dusk #dusk $PORTAL
If one machine can run both, would you still separate the roles?