💬 Want to contribute? Join us on Telegram: https://t.me/+DOylgFv1jyJlNzM0

Problem

backend/src/services/webhookService.ts sends webhooks with a single attempt (line 189: attempt_count, 1). If a delivery fails (timeout, 5xx response, network error), it is recorded in webhook_deliveries but never retried.

This means webhook subscribers can silently miss events with no way to recover other than re-polling.

Expected Behavior

Failed webhook deliveries should be retried with exponential backoff (e.g., 1m, 5m, 30m) up to a configurable max retry count.

Implementation Suggestion

1. Add a max_attempts column to webhook_subscriptions (default 5)
2. Add a scheduled job that queries webhook_deliveries for failed deliveries where attempt_count < max_attempts and delivered_at IS NULL
3. Retry with exponential backoff
4. Mark as permanently failed after max attempts

Files to Change

• backend/src/services/webhookService.ts — add retry logic
• backend/migrations/ — add migration for max_attempts column
• New: backend/src/services/webhookRetryScheduler.ts

Acceptance Criteria

• Failed deliveries are retried up to N times with exponential backoff
• webhook_deliveries.attempt_count is incremented on each retry
• Permanently failed deliveries are marked distinctly
• Retry scheduler runs on a configurable interval
• Admin endpoint to view retry queue status