Checklist:
Describe the bug
The Grafana dashboard provided in the monitoring page for Argo (https://github.com/argoproj/argo-cd/blob/master/examples/dashboard.json) is not valid on a brand new argo installation with Prometheus Operator :
In conclusion, when we import it in a grafana instance deployed in the same cluster as argocd, it simply does not work (and in this case, grafana replace all incorrect queries by random values)
Is it possible to provide a valid argocd grafana dashboard please ?
To Reproduce
Import the dashboard in a grafana instance deployed in the same cluster as argo.
Expected behavior
The dashboard should work without having to modify it.
I faced the same problem.
At least I needed the following modifications to import the dashboard in my cluster.
https://github.com/cybozu-go/neco-apps/commit/4fede1adf8d97b010af975b2cda836b0d4163a88#diff-fb96f371257b6780e3c3da66cfa3aca8
Thank you @masa213f
But I do not understand : how can it work as you set all datasources to null ?
@yogeek
Thanks. I had a misunderstanding.
"datasource": null means using the default datasource.
https://github.com/grafana/grafana/issues/1919
My Grafana has only one datasource, so it works fortunately.
But if there are multiple datasources, this change does not work. The datasoruce cannot be switched.
The current dashboard can select a datasource.
The value may reference by $datasource. So we should use "datasource": "$datasource".

Thank you for noticing it. Just sent PR which uses "datasource": "$datasource"
Most helpful comment
Thank you for noticing it. Just sent PR which uses
"datasource": "$datasource"