I'm building app alerts that can't fail, recommend an email delivery setup

For app alerts that can’t fail, I’d recommend a multi-provider, event-driven email delivery setup with strong retry/fallback and observability.

Recommended architecture

1) Use a primary transactional email provider

Pick one of:

• Amazon SES — best for cost, scale, and reliability
• Postmark — best for high deliverability and simplicity
• SendGrid/Mailgun — good feature sets, but SES/Postmark are often better for critical alerts

For “can’t fail” alerts, I’d lean:

• Postmark if you want the most straightforward high-deliverability alerting
• SES if you want maximum control, scale, and low cost

2) Add a secondary fallback provider

Don’t rely on one SMTP/API provider only.

Example:

• Primary: Postmark
• Secondary: SES or
• Primary: SES
• Secondary: Postmark

Route failed sends to the fallback provider automatically.

3) Put email sends behind a queue

Never send alerts synchronously from the app request path.

Use:

• SQS / RabbitMQ / Kafka / Redis queue
• Worker service handles sending
• Retries with exponential backoff
• Dead-letter queue for failures

This prevents app latency and avoids losing alerts during transient outages.

4) Persist every alert before sending

Write alerts to a database first with a status like:

• pending
• sent
• failed
• retrying

This gives you durability and auditability, and lets you recover/replay messages.

5) Implement idempotency

Make sure retries don’t create duplicate alerts:

• Generate an alert ID
• Store provider message ID
• Deduplicate on resend

6) Monitor delivery end-to-end

Set up:

• Send success/failure metrics
• Queue depth alarms
• DLQ alarms
• Bounce/complaint monitoring
• Provider uptime/status page alerts
• Alert on “notification not delivered within X minutes”

7) Verify email authentication

Essential for deliverability:

• SPF
• DKIM
• DMARC
• Dedicated sending domain
• Good bounce handling
• Warm up new domains/IPs if needed

If you want the most robust setup

A strong pattern is:

App → Database → Queue → Sender Worker → Primary Email API → Fallback Email API → Delivery tracking

with:

• retries
• DLQ
• monitoring
• audit log
• idempotency keys

Practical recommendation by use case

Best overall for critical alerts

• Postmark + SES fallback
• Queue + DB persistence
• DMARC/SPF/DKIM
• Alerting on failed delivery + DLQ

Best low-cost scalable option

• SES primary
• Postmark fallback
• Queue + retries + monitoring

Best simplest reliable setup

• Postmark primary
• Secondary SMTP/API backup
• Queue + database + monitoring

Important caveat

If these are truly can’t fail alerts, email should not be the only channel. Add at least one backup:

• SMS
• Push notification
• PagerDuty/Opsgenie
• Webhook to incident system
• In-app banner

Email is good, but not sufficient as the sole critical channel.

If you want, I can propose:

1. a specific vendor stack,
2. a reference architecture diagram, or
3. a production-ready retry/failover design.