fix(db): Clarify db.query.text and db.query.summary attributes

[Lms24]

• db.query.text: We previously specified this attribute to contain a full/raw query without parameterization. OTel requires the paramterized query.
• db.query.summary We specified this attribute to contain the paramterized query. OTel specifies a shortened operation/table "grouping".

This came up while reviewing getsentry/sentry-javascript#17961

[lcian]

LGTM, you will need to regenerate and reformat.
Also, example needs to be a string, we can change that but for the purpose of the PR just provide a single example as a string.

[Lms24]

+1 for being able to add more than one example. For now I removed the second example. Thanks for the review!

[cleptric]

@mjq and @Ahmed-Labs should take a look here as well.

[Ahmed-Labs]

Makes sense! We will probably have to update db attribute normalization logic because we currently expect a raw query as opposed to a parametrized one in db.query.text.

[Lms24]

Just to add here, JS SDK instrumentations have been sending the parametrized version in db.query.text ever since sending this attribute. I don't have the full picture if other SDKs handle this differently. If so, we need to be very mindful with this change in specification but it also means we're living with duality (in terms of semantics) already anyway.

[Lms24]

@Ahmed-Labs @cleptric should we move forward with this change?

[aldy505]

If I have this query:

INSERT INTO exec_test (id, name) VALUES (?, ?, ?, ?)

Should it become as it is, or:

INSERT INTO exec_test (id, name) VALUES (?)