As per the current OpenTelemetry semantic conventions, the span names we report are simply the database name. This simply isn't very useful.
While I intend to work on the specs themselves and possibly propose more useful options, we could also do our own thing and allow users to opt into other span name patterns via an API. At the extreme we could allow users to even provide a parameterized template which we'd evaluate (e.g. DB: {db} SQL: {sql}, but let's try to avoid log4j-like catastrophes please).
As per the current OpenTelemetry semantic conventions, the span names we report are simply the database name. This simply isn't very useful.
While I intend to work on the specs themselves and possibly propose more useful options, we could also do our own thing and allow users to opt into other span name patterns via an API. At the extreme we could allow users to even provide a parameterized template which we'd evaluate (e.g.
DB: {db} SQL: {sql}, but let's try to avoid log4j-like catastrophes please).