Best Practices + Troubleshooting

Best practices

• Make your handler idempotent; use job_id to dedupe
• Return 2xx within 30s; do heavy work asynchronously
• Log the event, signature status, and your own reference ID
• Use separate subscriptions per environment (sandbox vs prod)
• Monitor failure_count via List endpoint; unpause if needed

Troubleshooting

• Receiving duplicates: ensure idempotency on your side; duplicates are expected with retries
• Not receiving events: check subscription status (active vs paused), confirm URL, verify your endpoint responds with 2xx
• Signature validation failing: ensure you use the exact raw body and correct secret
• Auto-paused: fix your endpoint, then call Unpause; we do not auto-resume