Unable to send post requests to deployed rasa "anymore"

I have been using the following URL to send POST request to my deployed bot (using quick install) using postman:

http://domainname.com:5005/webhooks/rest/webhook

It was working just fine a few days back. Yesterday I made some changes to the deployment by adding a custom action image. Now I am getting a timeout error.

Could not send request

I am unable to understand what went wrong as I can’t site any warning/ error in any of the pods.

Even after removing the custom image from environment variables and values.yml, the issue persists.

On asking to describe the rasa-production pod, I can see the following warning:

Warning  Unhealthy  16m   kubelet            Liveness probe failed: Get 
"http://10.42.0.10:5005/": context deadline exceeded (Client.Timeout exceeded while 
awaiting headers)

Hi,

The rasa-production pod is not exposed outside of a Kubernetes cluster, by default, the Quick-Installation exposes Rasa X on port 80 (kubectl get ingress). The address that you use with port 5005 seems to be incorrect.

That you can see Liveness probe failed error for the rasa-production pod is quite normal at the very begging when the pod is starting, the important is to check the status if it’s running and ready.

The warning still persists after running the instance for days. So, it doesn’t seem like sth which will go away with time.

I didn’t fully understand what you’re suggesting here. I have used http://domainname.com:5005/webhooks/rest/webhook successfully in past.

I didn’t fully understand what you’re suggesting here. I have used http://domainname.com:5005/webhooks/rest/webhook successfully in past.

You mentioned that you use quick-install, I assume that you mean quick-install that is described here Quick-Installation. The quick-install script is using the rasa-x helm chart that by default doesn’t expose anything on 5005 port.

If you were able to access your deployed bot on 5005 port it means that you probably use a custom configuration that exposes the service on 5005 port.

Please check what version of the rasa-x helm chart is used currently and what version was used before your changes.

Additionally, it might be helpful if you can provide the comments/steps that you execute to update your deployment.