What

Removes the end-of-day health check that posted per-user metrics to the #customer-success Slack channel.

The feature was a single self-gating cron , runEndOfDayHealthCheck, dispatched hourly by EventBridge and gated to midnight UTC , that emitted three recap messages:

1. Daily summary , meetings + CRM events processed, by provider
2. Draft send rate per user , every active user's send rate (with a below-threshold call-out)
3. Automations processed per user , runs-per-user, per AOP

Why

We no longer want this daily recap going to customer-success.

Changes

• Delete apps/server/src/services/notifications/internal/health-checks.ts , all three report builders + the orchestrator (no other callers).
• cron/cron-task-registry.ts , remove the runEndOfDayHealthCheck entry and its per-UTC-date dedup guard.
• aws/lib/stacks/app-stack.ts , remove the hourly cron registration; tidy an adjacent comment that referenced it.
• services/notifications/types.ts , remove the daily_summary / low_send_rate_summary / automations_processed notification types, their union members, and the now-unused SimplifiedSourceSummary import.
• services/notifications/internal/slack-notifier.ts , remove the three rate-limit keys, the three Slack formatters, the isDailyRecap logic, and the customer-success channel routing branch.
• trpc/routes/admin-notifications.ts , remove the two admin test-send cases + the now-unused today local.
• services/analytics/draft-analytics.ts , remove getActiveUsersSimplifiedDraftSummary (the health check was its only caller).

Notes

• CEDAR_SLACK_CHANNELS.CUSTOMER_SUCCESS is retained , still used by the deal-audit step.
• Lower-level analytics helpers getActiveUsersDraftSummaryBySource / toSimplifiedSourceSummary are retained as general module API.
• No test edits required: the cron-registry wiring test derives task names from the registry source, so it stays green.
• @zero/server typecheck shows no new errors on any touched file.

🤖 Generated with Claude Code

Greptile Summary

The PR removes the end-of-day customer-success health check and its supporting analytics, cron, Slack formatting, routing, notification types, and server-side test cases.

• Removes the hourly infrastructure schedule and cron registry entry.
• Deletes the health-check implementation and its now-unused analytics wrapper.
• Removes recap payload types, Slack formatters, rate-limit keys, and customer-success routing.
• Removes server-side admin test cases, but leaves the corresponding mail playground controls and calls in place.

Confidence Score: 4/5

This needs the stale admin playground controls and calls removed before merging so the client and server notification contracts remain consistent.

The server no longer accepts two recap notification values that the retained mail playground still submits through rendered controls, causing typed-client incompatibility and runtime validation failures.

Files Needing Attention: apps/server/src/trpc/routes/admin-notifications.ts and apps/mail/app/(routes)/playground/notification-test.tsx

Important Files Changed