Skip to content
BulletinMail

Day-2 operations

Day-2 operations for a running BulletinMail instance.

Stub. Fill in as operational patterns emerge from running the reference instance. Below is the topic outline.

Routine

  • Daily: scan abuse@<apex> (manually until automated). Check Cloudflare Email Routing logs for anomalies.
  • Weekly: review DMARC aggregate reports landing at dmarc@<apex>.
  • Monthly: review per-tenant message volume against the configured daily limits; bump limits or follow up with the tenant as needed.

Incidents

  • Tenant being spammed / spamming: suspend via UPDATE tenants SET status='suspended' WHERE slug=?.
  • Bounce surge to a single domain: check Email Sending dashboard for IP reputation issues at that domain; coordinate with Cloudflare support.
  • Queue backlog: inspect dead-letter queue (wrangler queues consumer ...); diagnose root cause before redriving.

Backups and disaster recovery

  • D1 snapshots: schedule wrangler d1 export to R2 nightly.
  • R2 attachments: enable bucket lifecycle / replication per your tolerance.
  • DNS: keep an offsite copy of the zone file.

Capacity planning

  • 100 tenants × 5 groups × 50 members × 4 messages/week ≈ 500k sends/month — fits comfortably in Workers Paid + Email Sending baseline.
  • Re-evaluate when any of those numbers grows ~5x.

Decommissioning a tenant

  • Tenant requests data export.
  • Email all admins a 30-day notice.
  • After grace period: DELETE FROM tenants WHERE slug=? (cascades).
  • Free up the slug after 90 days (avoid surprise re-use).