metadata.json: re-add Puppet 8 support

[bastelfreak]

This was removed by mistake previously.

[wyardley]

@bastelfreak I came across this error in some other recent PRs (where with certain OS / RMQ versions, the user gets configured as rabbit vs. rabbit_cluster. It makes no sense to me, since the configs should still be properly defining it, but maybe a provider issue?

[bastelfreak]

puppet 8:

Failure/Error: expect(r.stdout).to match(%r!({cluster_name,<<"rabbit_cluster">>}|^Cluster name: rabbit_cluster$)!)
  expected "[{nodes,[{disc,['rabbit@centos7-64-puppet8']}]},\n {running_nodes,['rabbit@centos7-64-puppet8']},\n {cluster_name,<<\"rabbit@centos7-64-puppet8.example.com\">>},\n {partitions,[]}]\n" to match /({cluster_name,<<"rabbit_cluster">>}|^Cluster name: rabbit_cluster$)/
  Diff:
  @@ -1,4 +1,7 @@
  -/({cluster_name,<<"rabbit_cluster">>}|^Cluster name: rabbit_cluster$)/
  +[{nodes,[{disc,['rabbit@centos7-64-puppet8']}]},
  + {running_nodes,['rabbit@centos7-64-puppet8']},
  + {cluster_name,<<"rabbit@centos7-64-puppet8.example.com">>},
  + {partitions,[]}]

vs puppet 7:

centos7-64-puppet7.example.com 17:13:19$ rabbitmqctl -q cluster_status
  [{nodes,[{disc,['rabbit@centos7-64-puppet7']}]},
   {running_nodes,['rabbit@centos7-64-puppet7']},
   {cluster_name,<<"rabbit_cluster">>},
   {partitions,[]}]