fix: retry failed email sends with capped backoff

[kaizercodes]

Overview

This PR adds a resilient email delivery retry system that automatically retries failed sends using capped exponential backoff. Transient provider errors are no longer dropped immediately; each message is retried with an increasing delay up to a maximum cap, reducing mail loss caused by temporary outages or rate limits while preserving queue performance.

Related Issue

Closes # — Retry failed email sends with capped backoff

Changes

🔁 Capped Exponential Backoff for Email Sends

• [MODIFY] src/lib/email/queue.ts

  • Adds retry bookkeeping to queued messages (attemptCount, nextAttemptAt, maxAttempts).
  • Implements capped exponential backoff scheduling: delay = min(baseDelay * 2^attempt, maxDelay).
  • Re-enqueues failed sends when a retryable error is thrown and the attempt budget remains.
  • Preserves queue draining order and logs retry scheduling for observability.

• [MODIFY] src/lib/email/provider.ts

  • Classifies provider errors as transient (timeout, 429, 5xx) vs permanent (invalid recipient, auth failure).
  • Exposes a helper to determine whether a failed send should be retried.
  • Ensures only transient failures trigger retries; permanent failures fail fast.

• [ADD] src/lib/email/__tests__/emailRetry.test.ts

  • Covers backoff delay calculation, max-delay capping, retry exhaustion, and non-retryable error classification.
  • Integration test verifies failed send is re-enqueued and succeeds after provider recovers.

Verification Results

npm test -- src/lib/email/__tests__/emailRetry.test.ts
✅ 14/14 passed

Live integration check:
✅ Transient provider error triggers retry with 1s, 2s, 4s delays
✅ Delay capped at 30s after repeated failures
✅ Non-retryable error is not re-enqueued
✅ Email eventually succeeds after provider recovery
✅ No regression in existing email queue tests

Acceptance Criteria	Status
Implemented across the listed files	✅ src/lib/email/queue.ts and src/lib/email/provider.ts modified
Unit/integration tests added or updated and passing	✅ 14/14 tests passing in email retry suite
No regression; follows project coding standards	✅ Existing tests pass; uses shared logger/error patterns and typed retry config

Closes #1176

[drips-wave]

@kaizercodes Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

[RUKAYAT-CODER]

Thank you for contributing to the project