API and dashboard doesn't display unhealthy statuses

Hi,

when testing the healthchecks, its seem like the API (and thus the dashboard) doesn't display the proper status of services?

I simply stopped one container and the debug says, correctly:

level=warning msg="Health check still failing. Backend: \"web@docker\" URL: \"http://10.0.2.88:80\" Reason: HTTP request failed: Get http://10.0.2.88:80/nginx_status: dial tcp 10.0.2.88:80: connect: no route to host"

But the API response used from the dashboard reports all containers up:

   "serverStatus":{
      "http://10.0.2.84:80":"UP",
      "http://10.0.2.85:80":"UP",
      "http://10.0.2.88:80":"UP",
      "http://10.0.2.89:80":"UP"
   }

Fortunately it looks like no traffic is being routed to the node.
Starting the container again, I can see traffic coming in as normal.

Anyone else seeing something similar?

Hello @perosb,

There was a PR merged (https://github.com/containous/traefik/pull/5407) just recently for that bug.

This should be resolved in the next release.

1 Like