Additional bulk operations: app start/stop/restart, service instance delete

The bulk action framework (multi-select checkbox column + bulk bar + backend guid-list fan-out endpoints returning a per-item BulkResult) currently covers app delete, route delete, and route unmap. Candidates for extending it, roughly in priority order:

App bulk start / stop / restart
The most likely to see real use after delete. CF v3 start/stop are synchronous per app, so these behave like route unmap: final per-item counts in one response, no job polling needed. The apps lists already have the checkbox column and bulk bar, so each op needs a backend endpoint (same fan-out pattern as native_apps_bulk_writes.go) plus a bulkAction entry in the two apps list configs.

Service instance bulk delete
Managed service instance deletes return async CF jobs, so this would ride the same per-item job handoff + polling path as app/route bulk delete. The service instances list has no checkbox column yet, so this also pulls in list work.

Later / weaker candidates

• Service instance bulk unshare (synchronous, pairs with the existing bulk share)
• Bulk role removal on the users tables
• Org/space bulk delete — async jobs again, but destructive enough that it may deserve extra friction rather than a one-click bulk action

Doesn't fit the current contract
Operations needing a parameter beyond the guid list (e.g. bulk map-route-to-app needs a target app) break the {"guids": [...]} body shape and would need their own design.

Start by comparing the existing app, route, and route-unmap bulk actions with native_apps_bulk_writes.go and the two apps list configurations. Choose one candidate operation, then trace its endpoint, bulkAction entry, and— for service instances—the list checkbox and async job handoff. Done means the selected operation follows the existing per-item result or polling contract.