No More Silent Webhook Failures: Werk24 Now Emails You When a Callback Fails

Jul 29, 2026 · Written by Jochen MattesOriginalsprache: EnglischEnglische Version

Most Werk24 integrations receive their results through a webhook: once we finish reading a drawing, we POST the result to a callback URL you control. When that delivery works, you never think about it. The interesting question is what happens when it doesn't.

Endpoints go down during a deploy. A firewall rule changes. An authentication header expires. A new handler throws a 500 on an edge case. In all of these cases the result is ready on our side — but it has nowhere to land.

The failure used to be invisible

Until now, a failed delivery was effectively swallowed. We stopped the job — there was no reachable recipient to hand the result to — but nothing told you it had happened. No error, no bounce, no signal. Just a request that never produced a result.

If you had error tracking like Sentry wired into your callback handler, you might eventually catch it. But a handler that never receives the request can't report an error it never saw. And realistically, a webhook you just stood up for a new project is rarely instrumented on day one — monitoring tends to arrive later, after the integration is already carrying traffic.

So the first sign of trouble was often a support message a few days later: "we submitted a drawing and never got anything back." By then the trail was cold, and figuring out why the callback failed meant a round-trip through us.

Now the failure emails you

The moment a callback delivery fails, we send an email to your account's error contact. It names the request, explains in plain language why the delivery failed, tells you what to check next — and keeps the raw technical fields in a clearly fenced block at the bottom, so whoever debugs it has everything in one place.

Just as importantly, the notification is best-effort and completely out of the request's way. Sending it can never slow down, block, or interfere with the reading of your drawings. If the email itself can't be sent, your request behaves exactly as it did before.

What the email looks like

Here is a real example — a callback to an endpoint that had gone offline. The plain-language explanation comes first, then the diagnostics, so the two never blur together:

Subject: Werk24: callback delivery to your endpoint failed (ClientConnectorError)

Hello,

We were unable to deliver a callback from Werk24 to your webhook
endpoint, so processing of your request could not be completed.

The callback could not be delivered (ClientConnectorError). Your
endpoint appears to be unreachable, offline, or to have timed out.

What you can do
- Check that your webhook endpoint is online and reachable from the
  public internet.
- Confirm the callback URL is not blocked by a firewall or network
  policy.

Once your endpoint is reachable again, please resubmit the request.
If you need help, just reply to this email and our team will assist you.

--------------------------------------------------------------------
Technical details
--------------------------------------------------------------------
Account:       you@your-company.com
Callback URL:  https://api.your-company.com/werk24/callback
Request ID:    3f2a9c1e-7b40-4e1a-9c2d-8a6f0b1d4e55
Message:       ASK / ASK_SHEET_ANOMALIES (page 1)
Error:         ClientConnectorError
Time (UTC):    2026-07-29 14:32:07
--------------------------------------------------------------------

This is an automated message from Werk24.

The advice matches the failure

The explanation and the "what you can do" steps change depending on how the delivery failed, so you are pointed straight at the likely cause:

What happened	What we tell you
Unreachable or timed out	Your endpoint appears to be offline or to have timed out. Check that it is online, reachable from the public internet, and not blocked by a firewall.
HTTP 4xx – rejected	Your endpoint rejected the callback — the webhook may be misconfigured. Verify the callback URL, authentication headers, and expected payload format.
HTTP 5xx – server error	Your endpoint returned a server error and could not process the callback. Check your callback handler's logs — it errored while handling our request.
HTTP 429 – rate limited	Callbacks are arriving faster than your system accepts them. Raise the rate your endpoint accepts, or reply to the email so we can discuss the throughput you need.

Make sure it reaches the right person

The notification is sent to your account's error contact. If you have configured an error_email for your account, that is where it lands — so set it to whoever owns your integration (a shared team inbox works well). If we cannot determine an address, the alert still is not lost: it routes to Werk24 operations so nothing slips through unseen.

Nothing else about your integration changes. Successful callbacks behave exactly as before — you will only hear from us when a delivery fails, which is precisely when you would want to know.

Questions about your webhook setup? Just reply to any Werk24 notification and our team will help — or get in touch with us.