Daily CRM summary — Production Report

Field	Value
Branch	feature/daily-crm-summary
Date	2026-07-23
Status	Ready for GA after migrate + emails worker smoke (Critical/High audit items addressed)
Repos	SaaS-Backend, SaaS-Frontend, SaaS-Docs

---

Summary

At workspace Daily Reminder Time (task_reminder_time, default 09:00 local), crm:send-due-notifications sends mail-only CRM snapshots in addition to the existing task due digest:

User flag receive_all_users_daily_summary	Email
false	Personal summary only (skipped when no open CRM work)
true	User-wise team summary only (not both); roster includes only users with activity

Counts:

• Leads — open stages only (exclude is_won / is_lost)
• Tasks — open / in_progress / waiting (exclude completed / cancelled)
• Meetings — scheduled only; distinct host or attendee (creator-only not counted)

Aggregations run once per tenant per command tick. Meeting counts use SQL COUNT(DISTINCT meeting_id).

---

Deploy checklist

1. Deploy Backend + Frontend + Docs from feature/daily-crm-summary.
2. Migrate:

   php artisan migrate --force

   • users.receive_all_users_daily_summary (default false)
   • daily_summary_deliveries (unique tenant_id, user_id, digest_date, kind)
3. Confirm scheduler: crm:send-due-notifications every 5 minutes (withoutOverlapping, onOneServer).
4. Confirm php artisan queue:work --queue=emails (and queue:restart after deploy).
5. Shared cache driver required for onOneServer locks.
6. Configure Settings → Daily Reminder Time; flag managers with Receive all-users daily summary (visibility grant — prefer Owner/Admin).
7. Smoke:
   • Past reminder time → personal mail to unflagged user with open CRM work
   • Team mail to flagged user (only users with activity)
   • Second cron tick → no duplicate
   • Ledger queued → sent; stale queued (>45m) reclaimable (max 5 attempts)
8. Keep scheduler + emails workers healthy through the local reminder window (no midnight catch-up).

---

Monitoring

Signal	What to watch
daily_summary_deliveries.status = queued older than ~45 minutes	Worker down (auto-reclaim after stale window)
attempts >= 5 still failed	Cap reached; investigate SMTP/queue
crm:send-due-notifications duration	Should stay flat vs user count after aggregate-once fix
Emails queue depth	Backlog after reminder time

---

Audit remediation (2026-07-23)

Item	Status
Personal N× full-tenant scans	Fixed — summarizeUsers() once per tenant
Meeting PHP hydration	Fixed — SQL COUNT(DISTINCT) union
Stale queued / max attempts	Fixed — 45m reclaim, max 5
Team zero-activity noise	Fixed — summarizeActiveUsers()
Deploy / database / Users API docs	Fixed
Pest gaps	Fixed — markSent, suspended, empty, TZ, dedupe, stale reclaim
Playwright flag toggle	Added users.daily-summary-flag.spec.ts
RBAC duplicate Edit rows + privacy note	Fixed

---

Related docs

• Tenant settings (user)
• Tasks (user)
• Tenant RBAC (user)
• Tenant Users API
• Tenant notifications API
• Database architecture
• Tasks production
• Changelog