Speaking as a SRE/DevOps guy with almost 20 years of experience:

There are no solutions, only tradeoffs.

I'll give some examples:

Firm 1 had a single server running tron [0] for ALL scheduling of processes.

Pros: very easy to see what should run when and made audits a breeze

Cons: the central server died and it was a giant outage response to make sure that when the server came up it didn't start killing processes that should be running

Firm 2 used a management gui to create custom cron entries on each machine

Pros: each node had a local copy of the schedule and could keep going even if the central system died

Cons: each node had a local copy of the schedule which could "drift" from other nodes, a node could be forgotten etc

So, more generally, I agree that it's good to label things as "ok" so that we don't get into flame wars etc. That being said, the more important point is to say that if you are going to pick a strategy, do the work to support, build tooling and plan for outages related to that strategy.