fix: increase Grafana Prometheus datasource query timeout from 5s to 60s

The 5s queryTimeout caused dashboard panels to fail with "422: timeout
exceeded" errors when querying VictoriaMetrics, especially for range
queries over periods with high-cardinality pg_stat_statements data.

VictoriaMetrics already has -search.maxQueryDuration=30s, but Grafana's
per-request timeout header (5s) was overriding it.

https://claude.ai/code/session_01SzJxzZNQjDQphaHyaX3RU7

Author: claude

config/grafana/provisioning/datasources/datasources.yml

@@ -23,7 +23,7 @@ datasources:
     isDefault: true
     jsonData:
       scrapeInterval: '5s'
-      queryTimeout: '5s'
+      queryTimeout: '60s'
       timeInterval: '5s'
       httpMethod: 'POST'