Consistent naming: db.system to db.system.name, namespace constants, remove db from system-specific names

[lmolkova]

Fix #1581

Addresses DB-related portion of #608, replaces db-related things from #1711 and #1613.

1. Rename db.system to db.system.name. It gives us a good opportunity to clean up constants, batching multiple breaking changes together (see below)
2. Removes deprecated constants from new enum and updates existing constants to follow consistent naming
3. Removes db prefix from system-specific attributes following Demo - remove root namespace from system-specific attributes and metrics #1711:
   • some systems (e.g. redis are cache and database and also a messaging system)
   • not all operations that are done on the database are related to DB (control plane, auth, etc)
   • we're quite inconsistent in naming
4. Adjusts _ -> . where it makes sense

Merge requirement checklist

• CONTRIBUTING.md guidelines followed.
• Change log entry added, according to the guidelines in When to add a changelog entry.
  • If your PR does not need a change log, start the PR title with [chore]
• schema-next.yaml updated with changes to existing conventions.

[lmolkova]

ingres seems to be too generic - easy to mix with a word "ingress", but it's a trademark. I'm totally open to change it back to ingres

[lmolkova]

Dynamodb is a trademark, but also "amazon redshift" is a trademark.

Also, s3 is a trademark, but we use aws.s3 in other places.

I'm suggesting to keep all things AWS consistent and use aws prefix regardless of what trademark is registered for

[lmolkova]

All IBM databases have independent trademarks. E.g. db2

Still, given that other cloud services come with provider as a namespace, I'm suggesting ibm.*.

Maybe @gyliu513 or @mamwl1 would have some input from the IBM side? (it aligns with ibm.watsonx.ai) and we'd use ibm.mq as a messaging system name.

[lmolkova]

should we rename it to ms.sql or ms.sql.server ? wdyt @trask @jcocchi ?

[trask]

my initial inclination is not to prefix db2, informix and netezza with ibm, because they are standalone databases that exist outside of ibm cloud (I think still?)

and so feel more like cassandra, geode, etc which we don't prefix with apache

[lmolkova]

Not everything has a trademark or a unique name - e.g. IBM MQ would have to become ibm.mq regardless of where it's hosted. At this moment we'd become slightly inconsistent.

Also, we use cloud.provider=ibm_cloud, not ibm to talk about the cloud - not sure if we want to keep it though.

Here ibm is a company.

I feel there are two options:

1. prefix with a company name when a company provides multiple products (i.e. use ibm here) regardless of a trademark or cloud provider. Some company names are the same as product names (mongoDB, easticsearch is similar - it's already prefixed). Apache projects are similar - there is no ambiguity and there is no company that owns all of them.
2. use company prefix only when there is a potential ambiguity (ibm.mq) and then I'd probably keep spanner, cosmosdb, dynamodb, redshift. At the end of the day, you can host some of those on prem, run in emulators, and cloud provider prefix is not meaningful anyway

[trask]

does this work now that we are sort of considering this to point to a top-level namespace?

or maybe a better question, if we add oracle (database) specific attributes what namespace would they be under? (now that we also have oracle cloud related namespaces I think)

[lmolkova]

renamed to oracle.db, but it looks funky along with couchdb, mariadb, cosmosdb, dynamodb... Need to sleep on this.

It's the case where we have one company and may offerings oracle.db, oracle.nosql_db?, oracle.web_logic, oracle.jvm 😅, oracle cloud is being added as oracle_cloud

[trask]

these get tricky...

I think my biases are showing...

if it's a brand name I'm familiar with like informix, I prefer not prefixing it with ibm.

but in this case I don't mind

[lmolkova]

I'm not a big fan of having sap.hana and maxdb both being SAP databases but following different naming patterns