feat(attribute-values): Add counts to response and orderby

[wmak]

• This adds counts to the AttributeValues endpoint so its a part of the response and the default sort so we're getting the most common values
• Also updates the protos to support this

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit bf90686. Configure here.}

[cursor]

Cursor-based page token incompatible with new sort order

Medium Severity

The response page_token constructs a filter_offset cursor using attr_value > values[-1], which assumes results are sorted alphabetically by attr_value. The PR changed the primary sort to count() DESC, attr_value ASC, so values[-1] is no longer the alphabetically greatest value — it's just the last item in count-descending order. This makes the cursor semantically incorrect: it would skip values that belong on the next page or include values that don't. The offset-based path also doesn't account for the non-deterministic nature of count-based ordering across requests.

Additional Locations (1)

• snuba/web/rpc/v1/trace_item_attribute_values.py#L139-L143

 

^{Reviewed by Cursor Bugbot for commit bf90686. Configure here.}

[cursor]

Redundant distinct wrapper with GROUP BY clause

Low Severity

f.distinct() wrapping attr_value is now redundant because the new groupby=[column("attr_value")] already guarantees unique values. In ClickHouse, this generates SELECT DISTINCT ... GROUP BY ... where the DISTINCT is a no-op. Keeping both deduplication mechanisms is confusing and obscures the query's actual semantics — GROUP BY is the operative deduplication here (needed for count()).

 

^{Reviewed by Cursor Bugbot for commit bf90686. Configure here.}

[phacops]

The 2 cursor comments should be fixed before merging.