From a fresh install and Grafana, it runs from port 80 not 3000 #14276
Description
What is the issue?
If you install Grafana following the instructions on the Linkerd website, the linkage from the linkerd viz dashboard, http://localhost:50750/grafana is broken. Internally, if you check linkerd's web-web pod, you can see a log something like
2025/07/26 13:50:25 http: proxy error: read tcp 10.1.0.183:50374->10.107.231.9:3000: read: connection reset by peer
If I then check in with grafana, kubectl get services --namespace grafana I see
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
grafana ClusterIP 10.107.231.9 80/TCP 15h
From the default install, for me anyway, grafana is starting up on port 80, not 3000 as assumed by the instructions on the page,
linkerd viz install --set grafana.url=grafana.grafana:3000 \
| kubectl apply -f -
How can it be reproduced?
From a fresh Docker for Desktop, Kubernetes, linkerd, linkerd-viz and grafana install per the docs.
Logs, error output, etc
2025/07/26 13:50:25 http: proxy error: read tcp 10.1.0.183:50374->10.107.231.9:3000: read: connection reset by peer
output of linkerd check -o short
Status check results are √
Environment
docker-desktop
kubeadm, 1 node, v1.32.2
$ linkerd version
Client version: edge-25.7.5
Server version: edge-25.7.5
$ kubectl version
Client Version: v1.33.3
Kustomize Version: v5.6.0
Server Version: v1.32.2
MacOS 15.5
Possible solution
Setting the linkerd vis grafana port to what it's running internally, 80, with this command
linkerd viz install --set grafana.url=grafana.grafana:80 | kubectl apply -f -
Fixes it
Additional context
No response
Would you like to work on fixing this bug?
None