TIMEOUT_DNSLOOKUP webhook

[Mochics]

Hello,

I got a pretty strange error going on.

I've built a simple chatbot with some fulfillment intents and configured the fulfillment section with an https endpoint that is working correctly.

My server is composed by:

• A landing page. Containing the chat widget,
• An http fulfillment server. Containing the business logic of the fulfillment.
• An nginx. Serving the SSL certificate and redirecting to both of this services depending on the endpoint received. 

The thing is that, most of the time, the connection between Dialogflow and my webhook server is working as expected but sometimes the chatbot does not respond. Nginx is not receiving any call so my http server code is not being executed.

Looking closely to the chatbot history, if I check the logs of the "non-responded messages", it pops up an error with the fulfillment process:

I can't understand what's going on. Does it seem to be a problem with the DNS resolution by Google?

The fact that it only fails "sometimes" is driving me crazy... I'd appreciate any clue.

Thanks!

P.S. I'm leaving here a communication diagram in case it helps.

[Olusayo Akinlaja]

Hello, Mochics

I understand you are facing DNS lookup issues from your webhook. This type of issue could be generally transient due to temporary network issues. It may also be observed when requests are being throttled due to overload. 

Is this issue easily reproducible or it is something that randomly occur? If it randomly occurs, can you verify the workloads at such times?

[Mochics]

Hello,

It randomly occurs, but it can be reproduced pretty consistent (in less than 1 minute of sending messages, I can reproduce it). What do you mean by "verifying the workloads"?

Thanks.