Docker Community Forums

Share and learn in the Docker community.

UCP Dashboard shows No Data


Below is screenshot on the Docker UCP dashboard:

Nodes’ metrics show “No Data” …

ucp proxy remains healthy:

check its logs:

have ever tried to restart manager node’s Docker engine but in vain.
On each node ‘docker stats’ is fine.
On each manager node ‘docker node ls’ is fine.
Internal firewall is off.

How to fix the issue? Please advise.

I have a support case open for the same problem.

In my case, it’s because kubernetes DNS is not working.

The following is logged in ucp-metrics-inventory logs:

2019-04-03T02:08:14.845116456Z curl: (6) Could not resolve host: ucp-controller.kube-system.svc.cluster.local
2019-04-03T02:08:34.937494366Z curl: (6) Could not resolve host: ucp-controller.kube-system.svc.cluster.local

Try to run the following on your UCP manager nodes:

UCP_VERSION=$(docker service inspect ucp-agent --format ‘{{index .Spec.Labels “com.docker.ucp.version”}}’)

docker run --rm --network container:(docker ps -qf name=ucp-metrics-inventory) docker/ucp-dsinfo:{UCP_VERSION} dig ucp-controller.kube-system.svc.cluster.local

In my case, the DNS queries are timing out.

Support has not resolved it yet.

Thanks for the info.

FYI, I had the same issue and was able to work around it by just having it use the short DNS name, not the FQDN:

kubectl -n kube-system patch daemonset ucp-metrics -p '{"spec":{"template":{"spec":{"containers":[{"name": "ucp-metrics-inventory", "args": ["-c", "while :; do curl -k -Ss --cacert ca.pem --cert cert.pem --key key.pem --output /inventory/inventory.json https://ucp-controller/metricsdiscovery; sleep 15; done"]}]}}}}'

I’ve same issue in only one node. One node has no data message

ucp-controller.kube-system.svc.cluster.local resolves fine.