Skip to main content

Webhook retries

If your endpoint returns non-2xx or times out, HelaMesh retries with exponential backoff:

Attempt 1 — immediately
Attempt 2 — 2 seconds later
Attempt 3 — 4 seconds later
Attempt 4 — 8 seconds later
…
Attempt 8 — 256 seconds later
After 8 attempts → marked FAILED

FAILED webhooks can be re-driven manually from the dashboard: Webhooks → click Redrive.

Idempotency is your responsibility​

The same invoiceId may arrive more than once if you returned 2xx but HelaMesh didn't see it (network blip). Always check your DB for an existing record before applying the payment.

// Bad — double-applies on retry
await db.orders.update({
  where: { helamesh_invoice_id: event.invoiceId },
  data: { payment_status: 'paid' }
});

// Good — no-op on retry
await db.orders.update({
  where: {
    helamesh_invoice_id: event.invoiceId,
    payment_status: { not: 'paid' }
  },
  data: { payment_status: 'paid' }
});

Timeouts​

HelaMesh considers a delivery failed if your endpoint doesn't respond within 10 seconds. Keep your webhook handler fast — acknowledge the event immediately, then do fulfillment work asynchronously.