docs(rum): add RUM resources Best practices note

[michael-richey]

Summary

Cleanup PR for the RUM initiative. Adds a RUM resources subsection under Best practices documenting the intake limitations, partial-CRUD adaptations, and excluded resources for the new RUM resource types. No code changes.

Stacked on #723 (rum_config) — this is the top of the RUM stack.

What it documents

• Intake data is not migrated (events, analytics, replay sessions, playlist→session associations).
• rum_operations / rum_replay_playlists are static definitions; query fields are RUM query filters, not session ids.
• rum_permanent_retention_filters: PATCH-only configure (create delegates to update, delete is a no-op).
• rum_teams_ownership_mappings: update = delete+recreate; team_handle is stable (not remapped); soft dep on teams.
• rum_config: singleton; only enforced_application_tags is configurable.
• Excluded: rum_teams_ownership_rules (read-only), rum_retention_quotas (no list endpoint).

Privacy audit

Inspected the cumulative diff from main to this stack head across all new model/test/README files. All test fixtures use obviously synthetic values (app-src, app-dst, rf-1, rf-dst, ef-1, op-1, op-dst, sl-1, pl-1, m-1, feat-1, team-platform, rum.metric.count, checkout-flow, prod-replay, rum-config). No real UUIDs, emails, tokens, org ids, or customer data present.

Testing

• Full unit suite: 1530 passed, 5 pre-existing subprocess-test failures unchanged, 8 skipped.

The RUM stack (in merge order)

1. test(rum): pin rum_applications model behavior; document in README #714 — rum_applications prep (tests + README)
2. feat(rum): add rum_metrics resource #715 — rum_metrics
3. feat(rum): add rum_retention_filters resource #716 — rum_retention_filters
4. feat(rum): add rum_retention_filters_order resource #717 — rum_retention_filters_order
5. feat(rum): add rum_permanent_retention_filters resource #718 — rum_permanent_retention_filters
6. feat(rum): add rum_operations resource #719 — rum_operations
7. feat(rum): add rum_operation_strong_links resource #720 — rum_operation_strong_links
8. feat(rum): add rum_replay_playlists resource (shell only) #721 — rum_replay_playlists
9. feat(rum): add rum_teams_ownership_mappings resource #722 — rum_teams_ownership_mappings
10. feat(rum): add rum_config resource (singleton) #723 — rum_config
11. this PR — README cleanup

[michael-richey]

Nice README follow-up.

This section makes the RUM resource scope and limitations much clearer (especially intake-data boundaries, singleton/configure-only semantics, and excluded endpoints). I think this will reduce operator confusion significantly.

No issues from my side.

[Copilot]

Copilot review overview

🟡 Changes recommended

The permanent-retention-filter documentation incorrectly describes the entire endpoint set as PATCH-only despite GET support.

Get a fresh assessment by requesting another Copilot review.

Review effort: Balanced
Findings: 1

Open (1)

• Clarify that GET enumeration is supported alongside PATCH-only mutations · New

What changed in this PR

Adds RUM resource limitations and synchronization behavior to the README.

Changes:

• Documents intake-data exclusions and partial-CRUD behavior.
• Clarifies resource dependencies and unsupported RUM resources.

File	Description
README.md	Adds RUM best-practices guidance.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[michael-richey]

Fixed in 5cf20b4f. Wording corrected to: "reads use GET, but mutations are PATCH-only (configure cross_product_sampling — no POST/DELETE)" so users don't infer the resource cannot be imported.

[michael-richey]

Thanks for the README follow-up — this improves operator expectations significantly.

The clarifications around:

• app-id constraints for app-scoped resources,
• singleton/update-only behavior, and
• resources that skip source↔destination mapping

are especially helpful for reducing surprise during first sync runs.

No blocking concerns from my side on this docs update.