From dc8a22f8c27f3ac3157a1a84cfc8d1b5c4232656 Mon Sep 17 00:00:00 2001 From: Jian Wang Date: Mon, 2 Feb 2026 10:11:34 +0100 Subject: [PATCH 1/2] Document how to correct the grafana invite link Signed-off-by: Jian Wang --- docs/troubleshooting/monitoring.md | 34 ++++++++++++++++++- .../troubleshooting/monitoring.md | 32 +++++++++++++++++ .../troubleshooting/monitoring.md | 32 +++++++++++++++++ .../troubleshooting/monitoring.md | 32 +++++++++++++++++ 4 files changed, 129 insertions(+), 1 deletion(-) diff --git a/docs/troubleshooting/monitoring.md b/docs/troubleshooting/monitoring.md index d9ef8c5331..36b0567a6c 100644 --- a/docs/troubleshooting/monitoring.md +++ b/docs/troubleshooting/monitoring.md @@ -692,7 +692,7 @@ cattle-monitoring-system prometheus-kubevirt-rules 24s // is missi ### Root Cause -When KubeVirt is newly installed or upgraded, it generates a new ConfigMap object to store the configuration. A race condition occurs within the KubeVirt operator if the `rancher-monitoring-operator` ServiceAccount object is missing/not synced from the `cattle-monitoring-system` namespace during this process. Consequently, the ServiceMonitor configuration may be excluded from the resulting ConfigMap object. +When KubeVirt is newly installed or upgraded, it generates a new ConfigMap object to store the configuration. A race condition occurs within the KubeVirt operator if the `rancher-monitoring-operator` ServiceAccount object is missing/not synced from the `cattle-monitoring-system` namespace during this process. Consequently, the ServiceMonitor configuration may be excluded from the resulting ConfigMap object. During the upgrade process, KubeVirt may incorrectly determine the monitoring state. Once the ConfigMap object is generated, KubeVirt does not reconcile or regenerate it until the next upgrade, unless a manual trigger is performed. @@ -790,3 +790,35 @@ The workaround involves ensuring that the `rancher-monitoring-operator` ServiceA ### Related Issue [#9674](https://github.com/harvester/harvester/issues/9674) + +## The embedded grafana does not provid correct URL path for invite + +### Issue Description + +1. Be on Harvester Dashboard +1. Click call-to-action/button to take us to Grafana landing page (rancher-monitoring) +1. Login +1. Click the gear-icon/button -> configuration -> then select 'users' tab (x-ref screengrab) +1. Build a new user, no email +1. Click the copy invite +1. Notice that when pasted in the browser it is directing somewhere in-accessible by the user, for example: `http://localhost:3000/invite/3qq7sUO9n7MF3XfRH8YfDWR94pYhli` + +### Root Cause + +The `grafana` depolyment on Harvester is sitting behind the ingress controller. A prefix is appended to the URL automatically when user visits the grafana per guided steps from Harvester UI. + +However, the generated invite link is not sent to external directly, it is still a localhost based link and can't be accessed from external. + +### Workaround + +Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/` to the invite link and remove the `http://localhost:3000`. The final link is `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/invite/smVzjAUMSx0WKfl0giG1w0zEQ0d6ID` + +:::note + +This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +::: + +### Related Issue + +[#5298](https://github.com/harvester/harvester/issues/5298) diff --git a/versioned_docs/version-v1.5/troubleshooting/monitoring.md b/versioned_docs/version-v1.5/troubleshooting/monitoring.md index da1d4441f1..36b0567a6c 100644 --- a/versioned_docs/version-v1.5/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.5/troubleshooting/monitoring.md @@ -790,3 +790,35 @@ The workaround involves ensuring that the `rancher-monitoring-operator` ServiceA ### Related Issue [#9674](https://github.com/harvester/harvester/issues/9674) + +## The embedded grafana does not provid correct URL path for invite + +### Issue Description + +1. Be on Harvester Dashboard +1. Click call-to-action/button to take us to Grafana landing page (rancher-monitoring) +1. Login +1. Click the gear-icon/button -> configuration -> then select 'users' tab (x-ref screengrab) +1. Build a new user, no email +1. Click the copy invite +1. Notice that when pasted in the browser it is directing somewhere in-accessible by the user, for example: `http://localhost:3000/invite/3qq7sUO9n7MF3XfRH8YfDWR94pYhli` + +### Root Cause + +The `grafana` depolyment on Harvester is sitting behind the ingress controller. A prefix is appended to the URL automatically when user visits the grafana per guided steps from Harvester UI. + +However, the generated invite link is not sent to external directly, it is still a localhost based link and can't be accessed from external. + +### Workaround + +Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/` to the invite link and remove the `http://localhost:3000`. The final link is `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/invite/smVzjAUMSx0WKfl0giG1w0zEQ0d6ID` + +:::note + +This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +::: + +### Related Issue + +[#5298](https://github.com/harvester/harvester/issues/5298) diff --git a/versioned_docs/version-v1.6/troubleshooting/monitoring.md b/versioned_docs/version-v1.6/troubleshooting/monitoring.md index da1d4441f1..36b0567a6c 100644 --- a/versioned_docs/version-v1.6/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.6/troubleshooting/monitoring.md @@ -790,3 +790,35 @@ The workaround involves ensuring that the `rancher-monitoring-operator` ServiceA ### Related Issue [#9674](https://github.com/harvester/harvester/issues/9674) + +## The embedded grafana does not provid correct URL path for invite + +### Issue Description + +1. Be on Harvester Dashboard +1. Click call-to-action/button to take us to Grafana landing page (rancher-monitoring) +1. Login +1. Click the gear-icon/button -> configuration -> then select 'users' tab (x-ref screengrab) +1. Build a new user, no email +1. Click the copy invite +1. Notice that when pasted in the browser it is directing somewhere in-accessible by the user, for example: `http://localhost:3000/invite/3qq7sUO9n7MF3XfRH8YfDWR94pYhli` + +### Root Cause + +The `grafana` depolyment on Harvester is sitting behind the ingress controller. A prefix is appended to the URL automatically when user visits the grafana per guided steps from Harvester UI. + +However, the generated invite link is not sent to external directly, it is still a localhost based link and can't be accessed from external. + +### Workaround + +Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/` to the invite link and remove the `http://localhost:3000`. The final link is `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/invite/smVzjAUMSx0WKfl0giG1w0zEQ0d6ID` + +:::note + +This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +::: + +### Related Issue + +[#5298](https://github.com/harvester/harvester/issues/5298) diff --git a/versioned_docs/version-v1.7/troubleshooting/monitoring.md b/versioned_docs/version-v1.7/troubleshooting/monitoring.md index da1d4441f1..36b0567a6c 100644 --- a/versioned_docs/version-v1.7/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.7/troubleshooting/monitoring.md @@ -790,3 +790,35 @@ The workaround involves ensuring that the `rancher-monitoring-operator` ServiceA ### Related Issue [#9674](https://github.com/harvester/harvester/issues/9674) + +## The embedded grafana does not provid correct URL path for invite + +### Issue Description + +1. Be on Harvester Dashboard +1. Click call-to-action/button to take us to Grafana landing page (rancher-monitoring) +1. Login +1. Click the gear-icon/button -> configuration -> then select 'users' tab (x-ref screengrab) +1. Build a new user, no email +1. Click the copy invite +1. Notice that when pasted in the browser it is directing somewhere in-accessible by the user, for example: `http://localhost:3000/invite/3qq7sUO9n7MF3XfRH8YfDWR94pYhli` + +### Root Cause + +The `grafana` depolyment on Harvester is sitting behind the ingress controller. A prefix is appended to the URL automatically when user visits the grafana per guided steps from Harvester UI. + +However, the generated invite link is not sent to external directly, it is still a localhost based link and can't be accessed from external. + +### Workaround + +Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/` to the invite link and remove the `http://localhost:3000`. The final link is `https://10.115.16.193/api/v1/namespaces/cattle-monitoring-system/services/http:rancher-monitoring-grafana:80/proxy/invite/smVzjAUMSx0WKfl0giG1w0zEQ0d6ID` + +:::note + +This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +::: + +### Related Issue + +[#5298](https://github.com/harvester/harvester/issues/5298) From e9a7637102fd4ad2597bb34c89e96ad5e5561c0d Mon Sep 17 00:00:00 2001 From: Jian Wang Date: Fri, 6 Feb 2026 10:52:58 +0100 Subject: [PATCH 2/2] Update per review comments Signed-off-by: Jian Wang --- docs/troubleshooting/monitoring.md | 4 +++- versioned_docs/version-v1.5/troubleshooting/monitoring.md | 4 +++- versioned_docs/version-v1.6/troubleshooting/monitoring.md | 4 +++- versioned_docs/version-v1.7/troubleshooting/monitoring.md | 4 +++- 4 files changed, 12 insertions(+), 4 deletions(-) diff --git a/docs/troubleshooting/monitoring.md b/docs/troubleshooting/monitoring.md index 36b0567a6c..201dc253f3 100644 --- a/docs/troubleshooting/monitoring.md +++ b/docs/troubleshooting/monitoring.md @@ -815,7 +815,9 @@ Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle- :::note -This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. +1. This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +1. When the Harvester cluster is imported to `Rancher Manager`, recommend to use [Role-based Access Control](https://ranchermanager.docs.rancher.com/integrations-in-rancher/monitoring-and-alerting/rbac-for-monitoring) and disable the `grafana` password login. ::: diff --git a/versioned_docs/version-v1.5/troubleshooting/monitoring.md b/versioned_docs/version-v1.5/troubleshooting/monitoring.md index 36b0567a6c..201dc253f3 100644 --- a/versioned_docs/version-v1.5/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.5/troubleshooting/monitoring.md @@ -815,7 +815,9 @@ Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle- :::note -This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. +1. This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +1. When the Harvester cluster is imported to `Rancher Manager`, recommend to use [Role-based Access Control](https://ranchermanager.docs.rancher.com/integrations-in-rancher/monitoring-and-alerting/rbac-for-monitoring) and disable the `grafana` password login. ::: diff --git a/versioned_docs/version-v1.6/troubleshooting/monitoring.md b/versioned_docs/version-v1.6/troubleshooting/monitoring.md index 36b0567a6c..201dc253f3 100644 --- a/versioned_docs/version-v1.6/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.6/troubleshooting/monitoring.md @@ -815,7 +815,9 @@ Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle- :::note -This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. +1. This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +1. When the Harvester cluster is imported to `Rancher Manager`, recommend to use [Role-based Access Control](https://ranchermanager.docs.rancher.com/integrations-in-rancher/monitoring-and-alerting/rbac-for-monitoring) and disable the `grafana` password login. ::: diff --git a/versioned_docs/version-v1.7/troubleshooting/monitoring.md b/versioned_docs/version-v1.7/troubleshooting/monitoring.md index 36b0567a6c..201dc253f3 100644 --- a/versioned_docs/version-v1.7/troubleshooting/monitoring.md +++ b/versioned_docs/version-v1.7/troubleshooting/monitoring.md @@ -815,7 +815,9 @@ Manually append the prefix like `https://10.115.16.193/api/v1/namespaces/cattle- :::note -This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. +1. This workaround also applies to other `grafana` local link, the link can be accessed from external after patching the prefix. + +1. When the Harvester cluster is imported to `Rancher Manager`, recommend to use [Role-based Access Control](https://ranchermanager.docs.rancher.com/integrations-in-rancher/monitoring-and-alerting/rbac-for-monitoring) and disable the `grafana` password login. :::