Why does the Grafana dashboard of a DM cluster display `failed to fetch dashboard` if the cluster is deployed using TiUP v1.3.0 or v1.3.1?

This is a known bug of TiUP, which is fixed in TiUP v1.3.2. The following are two solutions to this issue:

  • Solution one:
    1. Upgrade TiUP to a later version using the command tiup update --self && tiup update dm.
    2. Scale in and then scale out Grafana nodes in the cluster to restart the Grafana service.
  • Solution two:
    1. Back up the deploy/grafana-$port/bin/public folder.
    2. Download the TiUP DM offline package and unpack it.
    3. Unpack the grafana-v4.0.3-**.tar.gz in the offline package.
    4. Replace the folder deploy/grafana-$port/bin/public with the public folder in grafana-v4.0.3-**.tar.gz.
    5. Execute tiup dm restart $cluster_name -R grafana to restart the Grafana service.