Webhook URL not considered valid, but it is

[Nick Bicknell]

Hello,

I needs some help with figuring out why my webhook URL suddenly became invalid starting about a week ago.  I don't want to give all the specifics here in a public form, as some are sensitive, but here's the basic issue: 

1. The URL can be curl'd (get or post) with no problem and returns a 200 response when you do so

2. The message received when adding the webhook is "Unable to connect to the specified the URL, please double check and try again."  Strangely, nothing shows up in our webserver logs that indicates we even received a request for validation, which is not true if we try to validate the http:// version of the URL, which redirects to https:// and displays a 302 error message (as expected).

3. This same URL and many like it worked fine a bit more than a week ago.  All webhook URLs are now failing (we provide a service that integrates with MailChimp and many of our customers use that, each with their own unique webhook URL).

4. We have an alternate domain pointing at the same backend servers (different front end) which works fine for any of the failed webhooks

So, I have two questions: 

1. Is there a way that I can test the URL for these webhooks that mimics the same calls made by the validator?

2. Is there a way that I can provide MailChimp Tech Support with the information to test these without publishing the URLs on a public forum...

Thanks!

Nick

[jesse]

We simply make a basic curl call and check for a 2xx response. I imagine if you try from outside your network/whitelisted stuff you'll quickly figure out what has started blocking us. There's not much point in us testing anything since it would just be exactly what you're doing in the webapp, and we already know what you're seeing there.


jesse

[Nick B.]

Thanks, Jesse.

I have tested the URL from multiple locations outside my networ, and can't get anything but a 200 response.  This is why it's puzzling.  It's also puzzling that we get no record in our server logs for these calls when they error out.

So, I think it would be good to have more visibility into the error, but since it is working on everything I try (inside and outside my network), I'm at a bit of a loss as to what to try next. Can I refer you to the ticket where I submitted the details of this to support?  

If not, then I suppose we'll have to remove our dependency on webhooks and move to periodic API calls to get recent changes. 

Thanks!

Nick

[jesse]

Yeah, we just can't connect to you. I'd check with your network folks and see what sort of blocks or ssl changes may have been made since things started not validating. I did verify that our libcurl/libnss/php versions haven't changed any time recently. Here's what the connection looks like:

* About to connect() to demo5.blahblahblah.com port 443 (#0)
*   Trying 50.28.*.*... * connected
* Connected to demo5.blahblahblah.com (50.28.*.*) port 443 (#0)
* Initializing NSS with certpath: /etc/pki/nssdb
* NSS error -12286
* Error in TLS handshake, trying SSLv3...
> GET /integrations/79/listener HTTP/1.1
User-Agent: MailChimp.com WebHook Validator
Host: demo5.blahblahblah.com
Accept: */*

* Connection died, retrying a fresh connect
* Closing connection #0
* Issue another request to this URL: 'https://demo5.blahblahblah.com/integrations/79/listener'
* About to connect() to demo5.blahblahblah.com port 443 (#0)
*   Trying 50.28.*.*... * connected
* Connected to demo5.blahblahblah.com (50.28.*.*) port 443 (#0)
* TLS disabled due to previous handshake failure
* NSS error -12286
* Closing connection #0
* SSL connect error
0



jesse

[Nick B.]

Thanks, Jesse.  That's very helpful.