After upgrading the Linux integration from 1.1.2 to 1.2.0 all the dashboards it manages are no longer working, which were moved to Linux node/. It appears that they require a cluster label. Was this added intentionally or was it a mistake?
How is the cluster label intended to be provided? I don’t see any configuration difference in the Grafana Agent static configuration.
I see that I can select “Selected” for cluster, which seems to clear the cluster label and show data on the dashboard until it is next refreshed, as it changes the cluster label back to All.
The logs dashboard with version 1.2.2 first displays multiple errors, where the cluster label is the cause. After deselecting the cluster, it resets all the other filters to All, allowing you to reselect the instance that you want to filter to. After every refresh, it will then just display one error, which is probably just the cluster label erroring, but doesn’t reset the any of the filters.
Hi, thanks for using Linux Integration. I have version 1.2.2 in my stack and don’t have any issues with agent reporting without cluster label as well. Does your error appear only when you enter logs dashboard directly from from folder/favorite dashboards. Does it appear when you open logs dashboard from another linux dashoboards by using dashboard links?
Hi vitaly. The list of errors appear on load when going from one of the other Linux Integration dashboards. Only the one error appears if you go to the logs using the starred list, like what appears on every refresh of the logs dashboard.
