RI-8196 Enrich BE analytics events with isProduction

[KrumTy]

What

Enriches existing backend analytics events with the connection's production status so we can answer the PRD's analytics questions ("did users run dangerous commands on prod?", "did Profiler start on prod?", "did bulk-delete confirm on prod?") without adding parallel FE PROD_CONFIRMATION_SHOWN events. Part of epic RI-6594 — Prod vs non-prod modes. Ticket: RI-8196.

Builds on RI-8177 (the isProduction field, 'true' | 'false' string convention from database.analytics.ts) and RI-8179 (DangerousCommandsProvider, an ACL-CAT-backed dangerous-command cache).

Events enriched

Module	Events	New fields
CLI	CLI_COMMAND_EXECUTED, CLI_COMMAND_ERROR_RECEIVED, CLI_CLUSTER_NODE_COMMAND_EXECUTED	isProduction, dangerous
Workbench	WORKBENCH_COMMAND_EXECUTED, WORKBENCH_COMMAND_ERROR_RECEIVED, SEARCH_* variants, WORKBENCH_INDEX_INFO_SUBMITTED	isProduction, dangerous (index info: isProduction only)
Bulk actions	BULK_ACTIONS_STARTED, BULK_ACTIONS_STOPPED, BULK_ACTIONS_SUCCEED, BULK_ACTIONS_FAILED	isProduction, dangerous (true for Delete/Unlink), confirmedThrough
Profiler	PROFILER_LOG_DOWNLOADED, PROFILER_LOG_DELETED, new PROFILER_STARTED	isProduction

Implementation notes

• Each analytics service injects DatabaseRepository and uses a private resolveIsProduction(...) helper, wrapped in try/catch (defaults to 'false' on lookup failure so analytics never blocks the request path).
• dangerous is computed in the caller (CLI/Workbench command executors) using DangerousCommandsProvider.getDangerousCommands(client) — the cache is per-database, so the cost is one ACL CAT round-trip per connection. Caller passes the boolean via additionalData; the analytics service stringifies it.
• For Bulk actions, dangerous is derived from BulkActionType (Delete/Unlink → 'true', Upload → 'false') since there's no command-level granularity.
• New optional confirmedThrough: 'standard' | 'type-to-confirm' field on CreateBulkActionDto, threaded through BulkAction → IBulkActionOverview → analytics. Defaults to null (for paths that don't go through a confirmation dialog, including all imports).
• DangerousCommandsProvider is now exported from DatabaseModule.
• New PROFILER_STARTED event fires from ProfilerService.addListenerForInstance after the listener attaches.

Out of scope (deferred)

Browser write events (key rename / delete / TTL change / value mutations, hash/list/set/zset/stream writes). The ticket's #5 scope item assumes existing BE analytics to enrich, but the codebase has none for those services — they're FE-tracked today. Adding them is a net-new module-wide telemetry effort and exceeds the "enrich existing emits" framing. RI-8184 will need to keep its FE confirmation event or get a follow-up BE ticket.

Testing

• Unit tests: 238 affected tests pass (cli-analytics, cli-business, workbench.analytics, workbench-commands.executor, bulk-actions.analytics, bulk-import.service, bulk-action, new profiler-analytics.service.spec.ts). Full API suite (3444 tests) green. New assertions cover:
  • isProduction='true' when the database is production-flagged
  • dangerous passthrough for command events and Delete vs Upload for bulk
  • confirmedThrough round-trips through the bulk-action DTO
  • isProduction='false' fallback when the repository lookup throws
• Lint: clean on all touched files.
• Type check: no new src/ errors (pre-existing test/ failures unchanged).
• Manual smoke (still to do before un-drafting): with a prod-flagged and a non-prod database connected, run GET foo and FLUSHDB from CLI/Workbench and confirm Segment payloads carry the correct isProduction and dangerous; trigger a bulk-delete with the FE confirmation dialog and confirm confirmedThrough='type-to-confirm'; open Profiler and confirm PROFILER_STARTED fires once.

🤖 Generated with Claude Code

[jit-ci]

🛡️ Jit Security Scan Results

✅ No security findings were detected in this PR

---

^{Security scan by Jit}

[github-actions]

Code Coverage - Backend unit tests

St.❔	Category	Percentage	Covered / Total
🟢	Statements	92.71%	15683/16916
🟡	Branches	74.77%	4953/6624
🟢	Functions	86.96%	2448/2815
🟢	Lines	92.55%	14992/16199

Test suite run success

3444 tests passing in 308 suites.

Report generated by 🧪jest coverage report action from a1c4135

[github-actions]

Code Coverage - Integration Tests

Status	Category	Percentage	Covered / Total
🟡	Statements	79.53%	17590/22117
🟡	Branches	62.08%	8102/13049
🟡	Functions	68.12%	2441/3583
🟡	Lines	79.1%	16545/20916