Appendix A — Grafana MCP Tool Reference Used in This Project

Tool	Purpose
query_prometheus	Instant/range PromQL queries for SLO metrics
query_prometheus_histogram	Latency percentile breakdowns
list_alert_groups	Check currently firing alerts
describe_infrastructure	Service topology and dependencies for root-cause scoping
query_loki_logs	LogQL queries for error signatures
query_loki_patterns	Detect common log patterns during correlation
tempo_traceql-search / tempo_get-trace	Pull exemplar traces through the request path
create_incident / add_activity_to_incident	Open and annotate a Grafana Incident
get_current_oncall_users / list_oncall_schedules	Identify who to page
generate_deeplink	Human-readable link back into Grafana for deeper digging
alerting_manage_rules	Read/adjust alert rules as part of remediation
create_annotation	Mark exactly when/where an issue began on a dashboard panel
get_panel_image	Render a dashboard panel as PNG for the control room UI and postmortem
get_incident / get_annotations / get_dashboard_summary	Postmortem assembly

See low-level-design.md for which agent owns which tools.