feat(spanv2): Compute hashed sentry.normalized_db_query and store it in attributes

[Ahmed-Labs]

Previously, relay wrote the hashed normalized description to span.group. We are now trying to move away from this vague attribute and instead use a more specific attribute to store hashes. In this case, the hash is used for span grouping for the insights query summary page (db queries get normalized into more general/non-specific queries and grouped by their hashes).

Note: bumping sentry conventions on a separate branch (and rebasing after it is merged)

[Dav1dde]

I see you also changed the conventions submodule, not sure if that was intentional since you already have #5439 open.

[Ahmed-Labs]

I see you also changed the conventions submodule, not sure if that was intentional since you already have #5439 open.

Oh thats just so relay conventions doesn’t break when I access the new attribute introduced by that conventions bump (after adding it to the defined consts). Was planning to rebase once that bump PR is merged

[jjbayer]

@Ahmed-Labs how does this hash written by Relay interact with the span grouping code in sentry? https://github.com/getsentry/sentry/blob/70575b5ffb0266231d0dc48445c70bf285770800/src/sentry/spans/grouping/strategy/base.py#L37. I might not understand all the intricacies of span grouping but would it make sense to not hash anything in Relay, and fully rely on sentry-side hashing / grouping instead?