Dashboard: Dashboard logs continuously showing http: TLS handshake error from 192.168.100.0:1234: EOF when dashboard is running in https

[kiich]

Dashboard version: image i'm using is - k8s.gcr.io/kubernetes-dashboard-amd64:v1.10.0
Kubernetes version: 1.9.3
Operating system: coreos

Have followed the recommended setup of creating my own certs and chain and everything that i believe is required.
I can access the dashboard in https using the config below, certs coming from secret and mounted in the usual way.

    spec:
      containers:
      - args:
        - --tls-cert-file=tls.crt
        - --tls-key-file=tls.key
        - --disable-skip
        image: k8s.gcr.io/kubernetes-dashboard-amd64:v1.10.0
[...]
        livenessProbe:
          failureThreshold: 3
          httpGet:
            path: /
            port: 8443
            scheme: HTTPS
[...]
        ports:
        - containerPort: 8443
          protocol: TCP

i expose this deployment via service on 443 -> 8443 and i can indeed see the dashboard and login via token.
But i've noticed on the POD log showing:

2018/11/15 22:04:24 Starting overwatch
2018/11/15 22:04:24 Using in-cluster config to connect to apiserver
2018/11/15 22:04:24 Using service account token for csrf signing
2018/11/15 22:04:24 No request provided. Skipping authorization
2018/11/15 22:04:24 Successful initial request to the apiserver, version: v1.9.3
2018/11/15 22:04:24 Generating JWE encryption key
2018/11/15 22:04:24 New synchronizer has been registered: kubernetes-dashboard-key-holder-kube-system. Starting
2018/11/15 22:04:24 Starting secret synchronizer for kubernetes-dashboard-key-holder in namespace kube-system
2018/11/15 22:04:24 Initializing JWE encryption key from synchronized object
2018/11/15 22:04:24 Creating in-cluster Heapster client
2018/11/15 22:04:24 Serving securely on HTTPS port: 8443
2018/11/15 22:04:24 Successful request to heapster
2018/11/15 22:04:31 http: TLS handshake error from 192.168.100.0:2456: EOF
2018/11/15 22:04:35 http: TLS handshake error from 192.168.101.0:60602: EOF
2018/11/15 22:04:35 http: TLS handshake error from 192.168.102.0:22088: EOF

and the same 3 lines looping, where 192.168.100|101|102 being in the range of my podCIDR.

I can not seem to find any reference to those messages in the code/github and I can guess whatever is trying to come in is coming in via http instead of https but i can't seem to find out who/what and how to fix this log entires.

thanks

[kiich]

Just to add the incoming IP does not seem to match any of the PODs i have running, at least from the output of kubectl get pods --all-namespaces and filtered on .status.podIP

in fact, what I should have said is, the message actually has the .0 ending, for example 192.168.1.0

i'll update the issue with this new IP info.

[kiich]

After finding similar issues for other Kubernetes related apps (in my case, this issue https://github.com/dexidp/dex/issues/992) with same message, it led me to look at healthcheck done by my ELB which indeed was initially set to TCP - changing this to SSL made the log to stop - Closing this issue!

Sorry for the confusion!