Grafana Dashboard
The AxonFlow Community Dashboard provides real-time visibility into your AI governance platform's security, performance, and policy enforcement metrics.
Quick Start
The dashboard is automatically loaded when you start AxonFlow with docker-compose:
docker compose up -d
# Access Grafana
open http://localhost:3000
# Login: admin / grafana_localdev456 (default)
Navigate to Dashboards → AxonFlow → AxonFlow Community to view the dashboard.
Dashboard Sections
Overview
Key metrics at a glance:
| Panel | Description |
|---|---|
| Total Requests | Total API requests processed |
| Blocked Requests | Requests blocked by security policies |
| Policy Evaluations | Number of policy evaluations performed |
| Success Rate | Percentage of allowed requests |
| P95 Pre-check Latency | 95th percentile gateway pre-check time |
| LLM Tokens Tracked | Total LLM tokens processed |
Security & Compliance
Real-time security monitoring panels:
| Panel | Type | Description |
|---|---|---|
| PII Detections by Type | Bar Chart | Breakdown of detected PII by type (Aadhaar, PAN, UPI, bank accounts, etc.) |
| Security Threats Blocked | Stat | Total blocked requests including SQL injection, PII leaks, and dangerous queries |
| Provider Distribution | Pie Chart | Distribution of Gateway Mode LLM audit calls by provider (OpenAI, Anthropic, Google, Azure, Ollama) |
| Active Policies Summary | Bar Chart | Policy enforcement activity - evaluations vs blocks vs PII detections |
Request Throughput
Traffic analysis panels:
| Panel | Description |
|---|---|
| Request Rate | Requests per second by status and component |
| Blocked Requests Rate | Blocked requests over time by component |
Latency
Performance monitoring panels:
| Panel | Description |
|---|---|
| Gateway Pre-check Latency | P50/P95/P99 for policy pre-check operations |
| Agent Request Latency | P50/P95/P99 for agent processing |
| Orchestrator Request Latency | P50/P95/P99 for orchestrator operations |
Policy Enforcement
Policy activity panels:
| Panel | Description |
|---|---|
| Policy Activity (Hourly) | Policy evaluations and blocked requests over time |
| Gateway Pre-check Results | Pie chart showing approved vs blocked ratio |
LLM Usage
LLM monitoring panels:
| Panel | Description |
|---|---|
| LLM Tokens (5m) | Token usage by type (input/output) |
| LLM Cost (Hourly) | Cost tracking by provider and model |
| LLM API Calls (5m) | API calls by provider and status |
MCP Connectors
Connector monitoring panels:
| Panel | Description |
|---|---|
| Connector Calls (5m) | Requests by connector type |
| Connector Latency | P50/P95 latency by connector |
| Connector Errors (5m) | Error tracking by connector and type |
Auto-Provisioning
The dashboard is automatically provisioned when Grafana starts. No manual import required.
Configuration files:
config/grafana/provisioning/datasources/prometheus.yml- Prometheus datasourceconfig/grafana/provisioning/dashboards/dashboards.yml- Dashboard provisioningconfig/grafana/dashboards/axonflow-community.json- Dashboard definition
Key Prometheus Metrics
These metrics power the dashboard panels:
Agent Metrics
| Metric | Labels | Description |
|---|---|---|
axonflow_agent_requests_total | status | Total requests by status |
axonflow_agent_blocked_requests_total | - | Blocked requests |
axonflow_agent_policy_evaluations_total | - | Policy evaluations |
axonflow_agent_request_duration_milliseconds_bucket | - | Request latency histogram |
Gateway Metrics
| Metric | Labels | Description |
|---|---|---|
axonflow_gateway_precheck_requests_total | approved | Pre-check requests |
axonflow_gateway_precheck_duration_milliseconds_bucket | - | Pre-check latency |
axonflow_gateway_rbi_pii_detected_total | pii_type, blocked | PII detections by type |
axonflow_gateway_llm_tokens_total | type | LLM token tracking |
axonflow_gateway_llm_cost_usd_total | provider, model | LLM cost tracking |
Orchestrator Metrics
| Metric | Labels | Description |
|---|---|---|
axonflow_orchestrator_requests_total | status | Total requests |
axonflow_orchestrator_blocked_requests_total | - | Blocked requests |
axonflow_orchestrator_policy_evaluations_total | - | Policy evaluations |
axonflow_orchestrator_llm_calls_total | provider, status | LLM calls by provider |
Connector Metrics
| Metric | Labels | Description |
|---|---|---|
axonflow_connector_calls_total | connector | Connector calls |
axonflow_connector_duration_milliseconds_bucket | connector | Connector latency |
axonflow_connector_errors_total | connector, error_type | Connector errors |
Customization
Adding Custom Panels
- Open the dashboard in Grafana
- Click Add → Visualization
- Configure your query using available Prometheus metrics
- Save the dashboard
Exporting Changes
To persist dashboard changes to source control:
- Click Dashboard settings (gear icon)
- Click JSON Model
- Copy the JSON
- Update
config/grafana/dashboards/axonflow-community.json
Troubleshooting
Dashboard Not Loading
# Check Grafana is running
docker compose ps grafana
# Check Grafana logs
docker compose logs grafana
# Verify provisioning files exist
ls -la config/grafana/provisioning/
No Data Showing
# Check Prometheus is scraping targets
open http://localhost:9090/targets
# Verify Agent is exposing metrics
curl http://localhost:8080/metrics
# Query Prometheus directly
curl http://localhost:9090/api/v1/query?query=axonflow_agent_requests_total
Dashboard Shows "No Data"
- Ensure you've made some API requests to generate metrics
- Check the time range selector (default: Last 1 hour)
- Verify the Prometheus datasource is configured correctly in Grafana
Next Steps
- Token Usage & Cost Tracking - Monitor LLM token usage
- CloudWatch Integration - AWS monitoring setup
- Audit Logging - Track all actions
Enterprise Dashboard
Enterprise deployments include additional dashboards with Executive Summary views, Cost Analytics, and Compliance reporting. Contact sales for details.