Re: should we enable log_checkpoints out of the box?

On Fri, Nov 5, 2021 at 10:50 AM Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> I think what we need is an orthogonal classification. "This FATAL here
> is routine; that ERROR there denotes a severe problem in the underlying
> OS". Additional levels won't help with that. Maybe adding the concept
> of "severity" or "criticality" to some messages would be useful to
> decide what to keep and what to discard.

It would be really nice if we could do that in some kind of sensible
way. I think it's hard to get agreement though because different
people are going to care about different things. In fact, that
consideration weighs heavily against trying to adopt any sort of
project policy at all: somebody's bound to think it sucks. Even people
who mostly like it will demand the ability to customize it. On the
other hand, I also can't endorse the current policy of generating a
paper trail long enough to reach Alpha Centauri, so maybe we ought to
try harder to reach some sort of consensus.

--
Robert Haas
EDB: http://www.enterprisedb.com