Update OM 2.0 with proposal#43 Created Timestamp syntax

[ArthurSens]

Updates OpenMetrics 2.0 with our discussions in the working group and prometheus/proposals#43

cc @Maniktherana

[krajorama]

Approved with some small corrections. Also as per WG decision, you'll need another approval.

[ArthurSens]

Approved with some small corrections. Also as per WG decision, you'll need another approval.

Thanks for clarifying about exemplars :)

[krajorama]

I wonder if @ct needs to be prohibited explicitly for gauges, gauge histograms and alike. The ABNF does allow it, so syntax doesn't restrict, needs to be a semantic rule.

[ArthurSens]

Comments addressed, thanks for the review!

I wonder if @ct needs to be prohibited explicitly for gauges, gauge histograms and alike. The ABNF does allow it, so syntax doesn't restrict, needs to be a semantic rule.

Do we mean that we need to be explicit in the Gauge/Gauge Histogram definition that they do not have created timestamps? The current definition doesn't say that created timestamps exist for them, so we can rely on the implicit information that those types don't have CT...?

[ArthurSens]

@krajorama, do you want to take another look or am I ok to merge?

[krajorama]

I'm ok to merge this with one comment about testing, see there.

I'd like to open a PR on top of this though to rename Created Value to Created Timestamp for clarity and make the doc consistent throughout. I feel that would read better, but is a separate question from changing _created to ct@.