On Sun, 9 Feb 2020 at 20:52, Tristan Colgate <[EMAIL PROTECTED]> wrote:
I'm not getting that distinction, this sounds like something internal to
Large numbers of features existing in pronql are *redundant" in that
label_replace only makes things Turing Complete across multiple recording
rule evaluations, so is far less powerful than PromQL in practical terms.
It would not be possible to do this feature via label_replace in a single
query for example, so the argument doesn't really make sense to me.
We have a few redundant functions for maths/historical reasons (e.g. the
addition of the exponentiation operator makes a few previously existing
functions questionable), and some to decrease support load (i.e. increase)
but these are all no more than minor syntactic sugar. What you're proposing
here is very different, as I already described.