OpenShift OVN-Kubernetes Benchmark MCP Server

# OVN-Kubernetes Latency Metrics Configuration # This file defines PromQL queries for collecting OVN-K latency metrics # The $interval placeholder will be replaced with '2m' by default metrics: # CNI Latency Metrics - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_node_cni_request_duration_seconds_bucket{command="ADD"}[$interval])) > 0) metricName: cni_request_add_latency_p99 unit: seconds type: cni_latency component: node description: "99th percentile latency for CNI ADD requests" - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_node_cni_request_duration_seconds_bucket{command="DEL"}[$interval])) > 0) metricName: cni_request_del_latency_p99 unit: seconds type: cni_latency component: node description: "99th percentile latency for CNI DEL requests" # Pod Annotation Latency - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_controller_pod_creation_latency_seconds_bucket[$interval]))) > 0 metricName: pod_annotation_latency_p99 unit: seconds type: pod_annotation_latency component: controller description: "99th percentile latency for pod annotation processing" # Pod Creation Latency Metrics - query: histogram_quantile(0.99, sum by(pod, le) (rate(ovnkube_controller_pod_lsp_created_port_binding_duration_seconds_bucket[$interval]))) metricName: pod_lsp_created_p99 unit: seconds type: pod_creation_latency component: controller description: "99th percentile latency for pod LSP creation to port binding" - query: histogram_quantile(0.99, sum by(pod, le) (rate(ovnkube_controller_pod_port_binding_port_binding_chassis_duration_seconds_bucket[$interval]))) metricName: pod_port_binding_p99 unit: seconds type: pod_creation_latency component: controller description: "99th percentile latency for pod port binding to chassis binding" - query: histogram_quantile(0.99, sum by(pod, le) (rate(ovnkube_controller_pod_port_binding_chassis_port_binding_up_duration_seconds_bucket[$interval]))) metricName: pod_port_binding_up_p99 unit: seconds type: pod_creation_latency component: controller description: "99th percentile latency for pod chassis binding to port up" - query: histogram_quantile(0.99, sum by(pod, le) (rate(ovnkube_controller_pod_first_seen_lsp_created_duration_seconds_bucket[$interval]))) metricName: pod_first_seen_lsp_created_p99 unit: seconds type: pod_creation_latency component: controller description: "99th percentile latency for pod first seen to LSP created" # Service Latency Metrics - query: sum by (pod) (rate(ovnkube_controller_sync_service_latency_seconds_sum[$interval])) / sum by (pod) (rate(ovnkube_controller_sync_service_latency_seconds_count[$interval])) metricName: sync_service_latency unit: seconds type: service_latency component: controller description: "Average service synchronization latency" - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_controller_sync_service_latency_seconds_bucket[$interval])) > 0) metricName: sync_service_latency_p99 unit: seconds type: service_latency component: controller description: "99th percentile service synchronization latency" # Network Configuration Application Metrics - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_controller_network_programming_duration_seconds_bucket[$interval])) > 0) metricName: apply_network_config_pod_duration_p99 unit: seconds type: apply_network_configuration component: controller description: "99th percentile latency for applying network configuration to pods" - query: histogram_quantile(0.99, sum by (pod, le) (rate(ovnkube_controller_network_programming_service_duration_seconds_bucket[$interval])) > 0) metricName: apply_network_config_service_duration_p99 unit: seconds type: apply_network_configuration component: controller description: "99th percentile latency for applying network configuration to services" # Configuration options config: default_interval: "2m" cache_expiry_minutes: 10 query_timeout_seconds: 600 max_concurrent_queries: 4