Currently, we use dittybopper to deploy grafana dashboards on a cluster where workloads are run, however it's only valid until life of the cluster itself.

We need to set up some infrastructure where Grafana is long running (outside of ephemeral OCP clusters). Initial use case is for Network Observability but perhaps could be extended to entire OCP QE Perf Scale team.

Discussion: https://coreos.slack.com/archives/GTHU29259/p1663689551637469