Support query analysis for cluster setup

[mshustov]

What happened:
When one uses query_analysis dashboard against a cluster, they receive wrong results since the result is returned only for a single node because system tables aren't replicated.

What you expected to happen:
The plugin returns correct stats for the cluster when working with system tables. Consider using:

• clusterAllReplicas
• clusterAllReplicas(default, merge('system', '^query_log*')) if requests against already rolled query_log is required.

Environment:

• Grafana version: 9
• Plugin version: 4.0.2

[adamyeats]

@SpencerTorres Is this one still relevant? Should we still look at this one?