Graph Lifecycle Notifications and Missed-Event Recovery Prompt

You are a senior platform engineer who builds Microsoft Teams change-notification pipelines and knows that a subscription that dies quietly is the worst kind of outage.

I will provide:
- The Graph resources we subscribe to (channel messages, chats, presence, membership) and the notificationUrl + lifecycleNotificationUrl setup
- How subscriptions are created, stored, and renewed today
- Any observed gaps: missed messages, expired subscriptions, reauthorizationRequired

Your job:

1. **Lifecycle event handling** — define handlers for each lifecycle event type (reauthorizationRequired, subscriptionRemoved, missed) and the exact action each must take.
2. **Renewal strategy** — design a renewal schedule that refreshes before expiration with margin, including how to recover a subscription that already expired versus one nearing expiry.
3. **Missed-event recovery** — on a "missed" signal, specify how to backfill using delta queries or a resync so no change is permanently lost, and how to dedupe replayed events.
4. **State and idempotency** — define the subscription registry (id, expiry, resource, deltaLink) and make downstream processing idempotent against replays.
5. **Observability** — list alerts for subscriptions approaching expiry, lifecycle events received, and validation-token failures so silent death is impossible.

Output as: (a) lifecycle-event handler table, (b) renewal + recovery flow described step by step, (c) the subscription registry schema, (d) the alerting checklist.

Assume every subscription will eventually expire or drop notifications — design the backfill path first and treat the live feed as best-effort, not a source of truth.