Plan requirement
| Subscription | Any plan |
| Access | Admin |
Webhooks fail quietly. Where to see what was sent, what came back, and the four usual causes of a failure.
Symptoms
- The other system receives nothing.
- It worked in testing and not in production.
- Some tickets arrive over there and some do not.
Look at the activity first
Zendesk records webhook calls with what was sent and what came back. That answers most questions immediately: whether it was called at all, and if so what the other end said.
No call recorded means the trigger did not fire, which is a trigger problem rather than a webhook one.
Resolution steps
1. The trigger did not match
Check the ticket events. If the trigger is not listed, the conditions are the cause.
2. Authentication failed
A 401 or 403 in the response. Credentials expired, or the token was rotated on the other side without anyone updating Zendesk. Extremely common after a few months.
3. The body was rejected
A 400. The other system did not like the shape of what arrived, often because a placeholder resolved to nothing and left invalid content.
4. The endpoint moved or is down
A timeout or a 404. Someone changed the address, or the receiving system is unavailable.
What happens after a failure
Zendesk retries a failed call for a while and then stops. Persistent failures can deactivate the webhook, and nothing tells you unless somebody is watching.
Nobody is alerted when a webhook stops. Where the integration matters, check it periodically. The usual way this is discovered is that a colleague notices records stopped appearing three weeks ago.
Comments
0 comments
Article is closed for comments.