Webhook signatures, raw bodies and replay

Fast answer

Start from the visible symptom and stable error identifier. Test the most likely layer first, use safe commands with expected output, apply one change, then prove recovery with an edge-ready route and a real external request.

Diagnostic decision path

1. Record the exact error ID, agent version, OS, route state and timestamp.
2. Confirm the local service independently on loopback.
3. Run oxaa routes list and oxaa diagnose --network.
4. Compare actual output with the documented expected state.
5. Apply the smallest safe correction.
6. Send a new external request and confirm the local response.
7. Escalate with reviewed redacted diagnostics if recovery fails.

Technical details

• raw request body before parsing
• provider secret
• signature header
• timestamp tolerance
• proxy header trust
• body truncation
• Oxaa replay versus provider redelivery
• single-use or expired credentials

Troubleshooting

Start with the local process, port and scheme. Then check authentication and clock, DNS, UDP/443, TCP/443 fallback, corporate proxy or TLS interception, route ownership and edge readiness. Use oxaa routes list, oxaa diagnose --network and stable error IDs; review diagnostic output before sharing it and remove secrets, bodies and private paths.

How traffic and Inspector data are handled

Oxaa's public edge terminates public HTTPS so it can resolve the exact hostname, identify the current authenticated route and forward the request. The session between the enrolled device and Oxaa is encrypted; the selected local service receives the request through that session.

Oxaa retains bounded account, route, usage, security, billing, support and diagnostic metadata required to operate and protect the service. Inspector body capture is a separate, explicit debugging action: the bounded request or response body copy shown by the local Inspector remains on the developer device and can be redacted, deleted or allowed to expire. Public traffic still has to be processed by the edge, so the accurate claim is local Inspector copies-not “Oxaa cannot see traffic.”

Cleanup and production handoff

Remove the external callback, preview URL or DNS binding; rotate test credentials; delete local captures; stop the route; and revoke the device or session when appropriate. Move production traffic to the deployed application, production ingress or event-delivery platform designed for that job.

What Oxaa is - and is not

Oxaa is development connectivity for a selected local HTTP or HTTPS service while the enrolled device and authenticated route are live. It is not application hosting, a general VPN, a forward proxy, permanent production deployment, raw TCP/UDP tunneling, arbitrary TLS passthrough, a CDN/WAF replacement or a production webhook delivery platform.