OpenShift OVN-Kubernetes Benchmark MCP Server

# OVS Metrics Configuration # File: metrics.yaml # This file defines PromQL queries and metadata for OVS monitoring apiVersion: v1 kind: metrics-config metadata: name: ovs-metrics description: "OVS Container and Process Metrics for OpenShift/Kubernetes" version: "1.0.0" # OVS Container and Process Metrics metrics: # CPU Usage Metrics - query: sum by(node) (irate(container_cpu_usage_seconds_total{id=~"/system.slice/ovs-vswitchd.service"}[5m])*100) metricName: ovs-vswitchd-cpu-usage unit: percent description: "OVS vswitchd CPU usage percentage per node" category: cpu threshold: warning: 70 critical: 85 - query: sum(irate(container_cpu_usage_seconds_total{id=~"/system.slice/ovsdb-server.service"}[2m]) * 100) by (node) metricName: ovsdb-server-cpu-usage unit: percent description: "OVSDB server CPU usage percentage per node" category: cpu threshold: warning: 50 critical: 75 # Memory Usage Metrics - query: ovs_db_process_resident_memory_bytes metricName: OVS_DB_MEMORY_SIZE_BYTES unit: bytes description: "OVS database process resident memory size" category: memory threshold: warning_mb: 500 critical_mb: 1000 - query: ovs_vswitchd_process_resident_memory_bytes metricName: OVS_VSWITCH_MEMORY_SIZE_BYTES unit: bytes description: "OVS vswitchd process resident memory size" category: memory threshold: warning_mb: 500 critical_mb: 1000 # Flow Table Metrics - query: ovs_vswitchd_dp_flows_total metricName: OVS_DATAPATH_FLOWS_TOTAL unit: flows description: "Total number of datapath flows in OVS" category: flows threshold: warning: 5000 critical: 20000 - query: ovs_vswitchd_bridge_flows_total{bridge="br-int"} metricName: OVS_BRIDGE_FLOWS_BR_INT unit: flows description: "Total flows in br-int bridge" category: flows threshold: warning: 10000 critical: 50000 - query: ovs_vswitchd_bridge_flows_total{bridge="br-ex"} metricName: OVS_BRIDGE_FLOWS_BR_EX unit: flows description: "Total flows in br-ex bridge" category: flows threshold: warning: 1000 critical: 5000 # Connection and Stream Metrics - query: sum(ovs_vswitchd_stream_open) metricName: OVS_STREAM_CONNECTIONS_OPEN unit: count description: "Total number of open OVS stream connections" category: connections - query: sum(ovs_vswitchd_rconn_overflow) metricName: OVS_RCONN_OVERFLOW_TOTAL unit: count description: "Total number of OVS remote connection overflows" category: connections threshold: warning: 100 critical: 1000 - query: sum(ovs_vswitchd_rconn_discarded) metricName: OVS_RCONN_DISCARDED_TOTAL unit: count description: "Total number of OVS remote connections discarded" category: connections threshold: warning: 50 critical: 500 # Additional OVS Performance Metrics - query: ovs_vswitchd_megaflow_cache_hits_total metricName: OVS_MEGAFLOW_CACHE_HITS unit: count description: "Total megaflow cache hits" category: performance - query: ovs_vswitchd_megaflow_cache_misses_total metricName: OVS_MEGAFLOW_CACHE_MISSES unit: count description: "Total megaflow cache misses" category: performance - query: rate(ovs_vswitchd_datapath_packets_total[5m]) metricName: OVS_DATAPATH_PACKET_RATE unit: packets_per_second description: "OVS datapath packet processing rate" category: performance - query: rate(ovs_vswitchd_datapath_errors_total[5m]) metricName: OVS_DATAPATH_ERROR_RATE unit: errors_per_second description: "OVS datapath error rate" category: performance threshold: warning: 10 critical: 100 # Network Interface Metrics (for OVS interfaces) - query: rate(container_network_receive_bytes_total{interface=~"ovs.*"}[5m]) metricName: OVS_INTERFACE_RX_BYTES_RATE unit: bytes_per_second description: "OVS interface receive bytes rate" category: network - query: rate(container_network_transmit_bytes_total{interface=~"ovs.*"}[5m]) metricName: OVS_INTERFACE_TX_BYTES_RATE unit: bytes_per_second description: "OVS interface transmit bytes rate" category: network - query: rate(container_network_receive_packets_dropped_total{interface=~"ovs.*"}[5m]) metricName: OVS_INTERFACE_RX_DROPPED_RATE unit: packets_per_second description: "OVS interface receive packets dropped rate" category: network threshold: warning: 10 critical: 100 # Global Configuration global: scrape_interval: 15s evaluation_interval: 15s query_timeout: 30s # Analysis Configuration analysis: enabled: true cpu_thresholds: warning_percent: 70 critical_percent: 85 memory_thresholds: warning_mb: 500 critical_mb: 1000 flow_thresholds: datapath_warning: 5000 datapath_critical: 20000 bridge_warning: 10000 bridge_critical: 50000 connection_thresholds: overflow_warning: 100 overflow_critical: 1000 discarded_warning: 50 discarded_critical: 500 # Alerting Rules (can be used with Prometheus AlertManager) alerting_rules: - alert: OVSHighCPUUsage expr: ovs_vswitchd_cpu_usage > 85 for: 5m labels: severity: critical component: ovs-vswitchd annotations: summary: "OVS vswitchd high CPU usage" description: "OVS vswitchd CPU usage is {{ $value }}% on {{ $labels.node }}" - alert: OVSHighMemoryUsage expr: ovs_db_process_resident_memory_bytes > 1073741824 # 1GB for: 5m labels: severity: critical component: ovsdb annotations: summary: "OVSDB high memory usage" description: "OVSDB memory usage is {{ $value | humanize }} on {{ $labels.instance }}" - alert: OVSHighFlowCount expr: ovs_vswitchd_dp_flows_total > 20000 for: 2m labels: severity: warning component: ovs-datapath annotations: summary: "OVS high datapath flow count" description: "OVS datapath has {{ $value }} flows on {{ $labels.instance }}" - alert: OVSConnectionOverflow expr: increase(ovs_vswitchd_rconn_overflow[5m]) > 100 for: 1m labels: severity: critical component: ovs-connections annotations: summary: "OVS connection overflow detected" description: "OVS connection overflow count increased by {{ $value }} in 5 minutes" # Dashboard Configuration (for Grafana integration) dashboards: - name: "OVS Performance Overview" panels: - title: "CPU Usage" type: graph metrics: - ovs-vswitchd-cpu-usage - ovsdb-server-cpu-usage - title: "Memory Usage" type: graph metrics: - OVS_DB_MEMORY_SIZE_BYTES - OVS_VSWITCH_MEMORY_SIZE_BYTES - title: "Flow Counts" type: stat metrics: - OVS_DATAPATH_FLOWS_TOTAL - OVS_BRIDGE_FLOWS_BR_INT - OVS_BRIDGE_FLOWS_BR_EX - title: "Connection Health" type: graph metrics: - OVS_STREAM_CONNECTIONS_OPEN - OVS_RCONN_OVERFLOW_TOTAL - OVS_RCONN_DISCARDED_TOTAL # Export Configuration export: formats: - json - yaml - prometheus retention: metrics: 30d analysis: 7d alerts: 90d