Prometheus Queries¶

You can query Prometheus directly at http://localhost:9092 to create custom visualizations or debug issues.

Useful Queries¶

Total API Operations:

rate(gateway_controller_api_operations_total[5m])

API Operations by Status:

rate(gateway_controller_api_operations_total[5m]) by (status)

Deployment Latency Percentiles:

histogram_quantile(0.95, rate(gateway_controller_deployment_latency_seconds_bucket[5m]))

xDS Clients Connected:

gateway_controller_xds_clients_connected

Database Operation Rate:

rate(gateway_controller_database_operations_total[5m])

Memory Usage:

gateway_controller_memory_bytes

HTTP Request Duration:

histogram_quantile(0.99, rate(gateway_controller_http_request_duration_seconds_bucket[5m]))

Request Rate:

rate(policy_engine_requests_total[5m])

Policy Execution Rate:

rate(policy_engine_policy_executions_total[5m])

Policy Execution Success Rate:

rate(policy_engine_policy_executions_total{status="success"}[5m]) /
rate(policy_engine_policy_executions_total[5m])

Average Request Duration:

rate(policy_engine_request_duration_seconds_sum[5m]) /
rate(policy_engine_request_duration_seconds_count[5m])

Active Streams:

policy_engine_active_streams

Error Rate:

rate(policy_engine_request_errors_total[5m])

Request Rate:

rate(envoy_http_internal_requests_total[5m])

Request Duration:

histogram_quantile(0.99, rate(envoy_http_request_duration_seconds_bucket[5m]))

Upstream 5xx Errors:

rate(envoy_http_upstream_rq_xx{envoy_response_flags="upstream_connect_fail"}[5m])