Grafana Dashboards

AegisGate Platform ships with 3 pre-built Grafana dashboards (34 panels total) that deploy automatically via the Helm chart’s ConfigMap.

Available Dashboards

Dashboard	Panels	Focus
`aegisgate-overview`	10	System health, request rates, error rates, latency, upstream status
`aegisgate-security`	12	Security operations: MTTD/MTTR, incidents, detection latency, scan pipeline
`aegisgate-mcp-agents`	12	MCP/agent monitoring: sessions, tool invocations, A2A failures, capability denials

Helm Deployment

The Helm chart includes a ConfigMap (grafana-dashboard-configmap.yaml) that deploys all 3 dashboards with the grafana_dashboard: "1" label for automatic discovery by Grafana’s sidecar provisioner.

# values.yaml
grafana:
  dashboards:
    enabled: true
  sidecar:
    dashboards:
      enabled: true
      label: grafana_dashboard

Dashboard Details

AegisGate Security (v4.4.0+)

12 panels for security operations teams:

Panel	Metric	Description
Security Scans by Result	`aegisgate_scan_results_total`	Count of scans by result (blocked, allowed, flagged)
Total Audit Events	`aegisgate_audit_events_total`	All audit events over time
Active Connections	`aegisgate_active_connections`	Current active proxy connections
Rate Limit Hits	`aegisgate_rate_limit_hits_total`	Rate limit enforcement events
Security Scans Over Time	`aegisgate_scans_total`	Scan throughput time series
MTTD (p50/p95/p99)	`aegisgate_mttd_seconds`	Mean Time To Detect histogram
MTTR (p50/p95)	`aegisgate_mttr_seconds`	Mean Time To Respond histogram
Incidents by Severity	`aegisgate_incidents_total`	Incidents grouped by severity level
Detection Latency by Framework	`aegisgate_detection_latency_seconds`	Per-framework detection timing
Incident Status Distribution	`aegisgate_incident_status`	Open vs. resolved vs. acknowledged
Scan Pipeline Duration	`aegisgate_scan_duration_seconds`	End-to-end scan pipeline timing
ML Shadow Predictions	`aegisgate_ml_predictions_total`	ML model predictions in shadow mode

AegisGate MCP/Agents (v4.4.0+)

12 panels for MCP and agent-to-agent monitoring:

Panel	Metric	Description
Active MCP Sessions	`aegisgate_mcp_active_sessions`	Current active MCP sessions
MCP Tool Invocations	`aegisgate_mcp_tool_invocations_total`	Tool call count by tool name
A2A Auth Failures	`aegisgate_a2a_auth_failures_total`	Agent-to-agent authentication failures
MCP Connections Over Time	`aegisgate_mcp_connections_total`	MCP connection throughput
Tool Invocations by Tool	`aegisgate_mcp_tool_invocations_total`	Breakdown by tool name
MCP Tool Results	`aegisgate_mcp_tool_results_total`	Success vs. error vs. denied
A2A Integrity Failures	`aegisgate_a2a_integrity_failures_total`	HMAC verification failures
A2A Capability Denials	`aegisgate_a2a_capability_denials_total`	Capability enforcement denials
A2A License Failures	`aegisgate_a2a_license_failures_total`	License-tier enforcement failures
Requests by Tier	`aegisgate_requests_by_tier_total`	Request volume by license tier
ML Inference Duration	`aegisgate_ml_inference_seconds`	ML model inference timing
Regex Scan Duration	`aegisgate_regex_scan_seconds`	Regex pattern matching timing

Prometheus Alerts

The Helm chart also deploys a PrometheusRule with 10 pre-configured alerts:

Alert	Severity	Trigger
HighScanBlockRate	warning	> 10% of scans blocked in 5 min
CriticalIncidentRate	critical	Any critical incident in 1 min
HighMTTD	warning	p95 MTTD > 5s for 10 min
HighMTTR	warning	p95 MTTR > 60s for 10 min
MCPSessionSpike	warning	> 100 active MCP sessions
A2AFailureSpike	warning	> 10 A2A auth failures in 5 min
RateLimitExceeded	info	Rate limit hits increasing
AuditLogGap	critical	No audit events for 5 min
MLDriftDetected	warning	ML model drift detected
UpstreamUnhealthy	critical	Upstream provider unhealthy

Manual Import

If not using Helm, import the JSON files directly into Grafana:

# Dashboard JSON files are in the Helm chart:
deploy/helm/aegisgate-platform/dashboards/aegisgate-overview.json
deploy/helm/aegisgate-platform/dashboards/aegisgate-security.json
deploy/helm/aegisgate-platform/dashboards/aegisgate-mcp-agents.json

In Grafana: Dashboards → Import → Upload JSON file.

Documentation