Today the HTTP APis are not versioned and not paginated:

• not versioned makes it hard to introduce or change functionality even with crdb's own internal client, because the APIs must behave well in mixed-version environments.

  By introducing versioning (e.g .in the URL) we make it easier to extend, substitute and deprecate functionality across multiple releases. IT also helps maintaining backward-compatibility by 3rd party tooling built by users.

• not paginated causes nodes to crash with OOM errors when the underlying resource is large, e.g. when retrieving log files, range/event logs, jobs tables etc.