cdc: UX consistency for changefeeds, like other database objects

[shermanCRL]

Changefeeds are “just” jobs, they are not an entity on the order of tables or indexes. While there is deeper work we could do for changefeeds to have descriptors, etc, we can add some UX in the short term to give changefeeds syntactic consistency with tables, indexes, etc. To be clear, changefeeds remain jobs, this is all UX.

• CREATE CHANGEFEED opt_name FOR TABLE...
  opt_name is new. Presumably stored in the job protobuf. We might follow the naming logic for indexes, where the name is optional and gets auto-created if not specified. If so, need a migration.
• ALTER CHANGEFEED <name> RENAME TO <name>
  Hmm, this already uses Job ID.
• SHOW CHANGEFEEDS
  An alias to SHOW CHANGEFEED JOBS, new column for name. Hide Job ID?
• DROP CHANGEFEED <name>
  Calls CANCEL JOB. Returns error if the specified name is not a changefeed job.
• DROP ALL CHANGEFEEDS
  Alias for CANCEL ALL CHANGEFEED JOBS
• PAUSE CHANGEFEED <name>
  Calls PAUSE JOB. Returns error if the specified name is not a changefeed job.
• PAUSE ALL CHANGEFEEDS
  Alias for PAUSE ALL CHANGEFEED JOBS
• RESUME CHANGEFEED <name>
  Calls RESUME JOB. Returns error if the specified name is not a changefeed job.
• RESUME ALL CHANGEFEEDS
  Alias for RESUME ALL CHANGEFEED JOBS
• SHOW CREATE CHANGEFEED <name>
  See also cdc: create changefeed from previous changefeed #69348

Proposed: for clarity, these new commands should only accept a name and not a job_id. If you want to operate with job_id's, use the existing syntax (which we will not deprecate). Alternatively, can we handle either a name or job_id without ambiguity or surprises? Or, should we not bother with their naming part and just use the following sugar with job_id?

We’d want to take care to do this in a way that we won’t regret in the future, by painting ourselves into a syntactic or semantic corner. Likely low-risk.

Jira issue: CRDB-22346

Epic CRDB-25041

[blathers-crl]

cc @cockroachdb/cdc

[surahman]

I am trying to get back into contributing to OSS with CRDB. Golang (😄) + DB (😄) = 😁

Could I please get some context on this issue - what is the plan of action here? I have been out of the CRDB codebase for a while. Is this issue time critical?