[rendering] Enum values not sufficiently separate from attribute footnotes

[dyladan]

Area(s)

area:other

What happened?

Description

When markdown is rendered, the attributes table is followed by footnotes, then by tables which list values for enums. Those enum tables are not sufficiently separated from the footnotes, and they appear to be a part of the footnotes when reading. For example in HTTP client semconv the error.type table of possible values appears to be part of footnote [14] even though it is not.

Example

Below is an example taken from HTTP client semconv that shows what I'm talking about

View example markdown

**[14]:** The `url.template` MUST have low cardinality. It is not usually available on HTTP clients, but may be known by the application or specialized HTTP instrumentation.

The following attributes can be important for making sampling decisions
and SHOULD be provided **at span creation time** (if provided at all):

* [`http.request.method`](/docs/attributes-registry/http.md)
* [`server.address`](/docs/attributes-registry/server.md)
* [`server.port`](/docs/attributes-registry/server.md)
* [`url.full`](/docs/attributes-registry/url.md)

`error.type` has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

| Value  | Description | Stability |
|---|---|---|
| `_OTHER` | A fallback error value to be used when the instrumentation doesn't define a custom value. | ![Stable](https://img.shields.io/badge/-stable-lightgreen) |

Above example rendered

[14]: The url.template MUST have low cardinality. It is not usually available on HTTP clients, but may be known by the application or specialized HTTP instrumentation.
The following attributes can be important for making sampling decisions
and SHOULD be provided at span creation time (if provided at all):

• http.request.method
• server.address
• server.port
• url.full

error.type has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

Value	Description	Stability
_OTHER	A fallback error value to be used when the instrumentation doesn't define a custom value.

What I'm asking for is some visual separation between these elements to make it more obvious that you have moved onto a separate section like this:

View example of suggested improved rendering (heading)

[14]: The url.template MUST have low cardinality. It is not usually available on HTTP clients, but may be known by the application or specialized HTTP instrumentation.
The following attributes can be important for making sampling decisions
and SHOULD be provided at span creation time (if provided at all):

• http.request.method
• server.address
• server.port
• url.full

Enum Values

error.type has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

Value	Description	Stability
_OTHER	A fallback error value to be used when the instrumentation doesn't define a custom value.

View example of suggested improved rendering (horizontal line separator)

[14]: The url.template MUST have low cardinality. It is not usually available on HTTP clients, but may be known by the application or specialized HTTP instrumentation.

The following attributes can be important for making sampling decisions
and SHOULD be provided at span creation time (if provided at all):

• http.request.method
• server.address
• server.port
• url.full

---

error.type has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

| Value | Description | Stability |
< |---|---|---|
| _OTHER | A fallback error value to be used when the instrumentation doesn't define a custom value. | |

Semantic convention version

v1.28.0

Additional context

No response

[dyladan]

I created two draft PRs to better show what I'm asking for. If there in consensus on one of them i'll make it a real PR

[trask]

Another option could be to add make the enum values into a footnote on the attribute type, e.g.

Attribute	Type	Description	Examples	Stability
http.connection.state	string [1]	State of the HTTP connection in the HTTP connection pool.	active; idle

[1]: http.connection.state has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

Value	Description	Stability
active	active state.
idle	idle state.

[dyladan]

Another option could be to add make the enum values into a footnote on the attribute type, e.g.

Attribute Type Description Examples Stability
http.connection.state string [1] State of the HTTP connection in the HTTP connection pool. active; idle
[1]: http.connection.state has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

Value Description Stability
active active state.
idle idle state.

This solves another issue which is that the attribute table doesn't say anything about a list of acceptable values. If you look at the semconv I linked originally, nowhere does it say in the attribute table that error.type, http.request.method, or any other enum, is an enum that has specific allowable values. If you don't happen to scroll down you might miss that entirely.

edit: as long as we're considering larger changes, i'd also suggest to call it "extensible enum" instead of "string" for type.