# services/monitoring/grafana-dashboard-network.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: grafana-dashboard-network
  labels:
    grafana_dashboard: "1"
data:
  atlas-network.json: |
    {
      "uid": "atlas-network",
      "title": "Atlas Network",
      "folderUid": "atlas-internal",
      "editable": true,
      "panels": [
        {
          "id": 1,
          "type": "stat",
          "title": "Ingress Success Rate (5m)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 6,
            "x": 0,
            "y": 0
          },
          "targets": [
            {
              "expr": "(sum(rate(traefik_entrypoint_requests_total{code!~\"5..\"}[5m]))) / clamp_min(sum(rate(traefik_entrypoint_requests_total[5m])), 1)",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "red",
                    "value": null
                  },
                  {
                    "color": "orange",
                    "value": 0.995
                  },
                  {
                    "color": "yellow",
                    "value": 0.999
                  },
                  {
                    "color": "green",
                    "value": 0.9995
                  }
                ]
              },
              "unit": "percentunit",
              "custom": {
                "displayMode": "auto"
              },
              "decimals": 2
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 2,
          "type": "stat",
          "title": "Error Budget Burn (1h)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 6,
            "x": 6,
            "y": 0
          },
          "targets": [
            {
              "expr": "(1 - ((sum(rate(traefik_entrypoint_requests_total{code!~\"5..\"}[1h]))) / clamp_min(sum(rate(traefik_entrypoint_requests_total[1h])), 1))) / 0.0010000000000000009",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "green",
                    "value": null
                  },
                  {
                    "color": "yellow",
                    "value": 1
                  },
                  {
                    "color": "orange",
                    "value": 2
                  },
                  {
                    "color": "red",
                    "value": 4
                  }
                ]
              },
              "unit": "none",
              "custom": {
                "displayMode": "auto"
              },
              "decimals": 2
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 3,
          "type": "stat",
          "title": "Error Budget Burn (6h)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 6,
            "x": 12,
            "y": 0
          },
          "targets": [
            {
              "expr": "(1 - ((sum(rate(traefik_entrypoint_requests_total{code!~\"5..\"}[6h]))) / clamp_min(sum(rate(traefik_entrypoint_requests_total[6h])), 1))) / 0.0010000000000000009",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "green",
                    "value": null
                  },
                  {
                    "color": "yellow",
                    "value": 1
                  },
                  {
                    "color": "orange",
                    "value": 2
                  },
                  {
                    "color": "red",
                    "value": 4
                  }
                ]
              },
              "unit": "none",
              "custom": {
                "displayMode": "auto"
              },
              "decimals": 2
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 4,
          "type": "stat",
          "title": "Edge P99 Latency (ms)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 6,
            "x": 18,
            "y": 0
          },
          "targets": [
            {
              "expr": "histogram_quantile(0.99, sum by (le) (rate(traefik_entrypoint_request_duration_seconds_bucket[5m]))) * 1000",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "green",
                    "value": null
                  },
                  {
                    "color": "yellow",
                    "value": 200
                  },
                  {
                    "color": "orange",
                    "value": 350
                  },
                  {
                    "color": "red",
                    "value": 500
                  }
                ]
              },
              "unit": "ms",
              "custom": {
                "displayMode": "auto"
              },
              "decimals": 1
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 5,
          "type": "stat",
          "title": "Ingress Traffic",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 8,
            "x": 0,
            "y": 4
          },
          "targets": [
            {
              "expr": "sum(rate(node_network_receive_bytes_total{device!~\"lo|cni.*|veth.*|flannel.*|docker.*|virbr.*|vxlan.*|wg.*\"}[5m])) or on() vector(0)",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "rgba(115, 115, 115, 1)",
                    "value": null
                  },
                  {
                    "color": "green",
                    "value": 1
                  }
                ]
              },
              "unit": "Bps",
              "custom": {
                "displayMode": "auto"
              }
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 6,
          "type": "stat",
          "title": "Egress Traffic",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 8,
            "x": 8,
            "y": 4
          },
          "targets": [
            {
              "expr": "sum(rate(node_network_transmit_bytes_total{device!~\"lo|cni.*|veth.*|flannel.*|docker.*|virbr.*|vxlan.*|wg.*\"}[5m])) or on() vector(0)",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "rgba(115, 115, 115, 1)",
                    "value": null
                  },
                  {
                    "color": "green",
                    "value": 1
                  }
                ]
              },
              "unit": "Bps",
              "custom": {
                "displayMode": "auto"
              }
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 7,
          "type": "stat",
          "title": "Intra-Cluster Traffic",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 4,
            "w": 8,
            "x": 16,
            "y": 4
          },
          "targets": [
            {
              "expr": "sum(rate(container_network_receive_bytes_total{namespace!=\"traefik\",pod!=\"\"}[5m]) + rate(container_network_transmit_bytes_total{namespace!=\"traefik\",pod!=\"\"}[5m])) or on() vector(0)",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "color": {
                "mode": "thresholds"
              },
              "mappings": [],
              "thresholds": {
                "mode": "absolute",
                "steps": [
                  {
                    "color": "rgba(115, 115, 115, 1)",
                    "value": null
                  },
                  {
                    "color": "green",
                    "value": 1
                  }
                ]
              },
              "unit": "Bps",
              "custom": {
                "displayMode": "auto"
              }
            },
            "overrides": []
          },
          "options": {
            "colorMode": "value",
            "graphMode": "area",
            "justifyMode": "center",
            "reduceOptions": {
              "calcs": [
                "lastNotNull"
              ],
              "fields": "",
              "values": false
            },
            "textMode": "value"
          }
        },
        {
          "id": 8,
          "type": "timeseries",
          "title": "Per-Node Throughput",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 8,
            "w": 24,
            "x": 0,
            "y": 8
          },
          "targets": [
            {
              "expr": "avg by (node) ((sum(rate(node_network_transmit_bytes_total{device!~\"lo|cni.*|veth.*|flannel.*|docker.*|virbr.*|vxlan.*|wg.*\"}[5m])) or on() vector(0) + sum(rate(node_network_receive_bytes_total{device!~\"lo|cni.*|veth.*|flannel.*|docker.*|virbr.*|vxlan.*|wg.*\"}[5m])) or on() vector(0)) * on(instance) group_left(node) label_replace(node_uname_info{nodename!=\"\"}, \"node\", \"$1\", \"nodename\", \"(.*)\"))",
              "refId": "A",
              "legendFormat": "{{node}}"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "unit": "Bps"
            },
            "overrides": []
          },
          "options": {
            "legend": {
              "displayMode": "table",
              "placement": "right"
            },
            "tooltip": {
              "mode": "multi"
            }
          }
        },
        {
          "id": 9,
          "type": "table",
          "title": "Top Namespaces",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 9,
            "w": 12,
            "x": 0,
            "y": 16
          },
          "targets": [
            {
              "expr": "topk(10, sum(rate(container_network_transmit_bytes_total{namespace!=\"\"}[5m]) + rate(container_network_receive_bytes_total{namespace!=\"\"}[5m])) by (namespace))",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "unit": "Bps"
            },
            "overrides": []
          },
          "options": {
            "showHeader": true
          },
          "transformations": [
            {
              "id": "labelsToFields",
              "options": {}
            }
          ]
        },
        {
          "id": 10,
          "type": "table",
          "title": "Top Pods",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 9,
            "w": 12,
            "x": 12,
            "y": 16
          },
          "targets": [
            {
              "expr": "topk(10, sum(rate(container_network_transmit_bytes_total{pod!=\"\"}[5m]) + rate(container_network_receive_bytes_total{pod!=\"\"}[5m])) by (namespace,pod))",
              "refId": "A"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "unit": "Bps"
            },
            "overrides": []
          },
          "options": {
            "showHeader": true
          },
          "transformations": [
            {
              "id": "labelsToFields",
              "options": {}
            }
          ]
        },
        {
          "id": 11,
          "type": "timeseries",
          "title": "Traefik Routers (req/s)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 9,
            "w": 12,
            "x": 0,
            "y": 25
          },
          "targets": [
            {
              "expr": "topk(10, sum by (router) (rate(traefik_router_requests_total[5m])))",
              "refId": "A",
              "legendFormat": "{{router}}"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "unit": "req/s"
            },
            "overrides": []
          },
          "options": {
            "legend": {
              "displayMode": "table",
              "placement": "right"
            },
            "tooltip": {
              "mode": "multi"
            }
          }
        },
        {
          "id": 12,
          "type": "timeseries",
          "title": "Traefik Entrypoints (req/s)",
          "datasource": {
            "type": "prometheus",
            "uid": "atlas-vm"
          },
          "gridPos": {
            "h": 9,
            "w": 12,
            "x": 12,
            "y": 25
          },
          "targets": [
            {
              "expr": "sum by (entrypoint) (rate(traefik_entrypoint_requests_total[5m]))",
              "refId": "A",
              "legendFormat": "{{entrypoint}}"
            }
          ],
          "fieldConfig": {
            "defaults": {
              "unit": "req/s"
            },
            "overrides": []
          },
          "options": {
            "legend": {
              "displayMode": "table",
              "placement": "right"
            },
            "tooltip": {
              "mode": "multi"
            }
          }
        }
      ],
      "time": {
        "from": "now-12h",
        "to": "now"
      },
      "annotations": {
        "list": []
      },
      "schemaVersion": 39,
      "style": "dark",
      "tags": [
        "atlas",
        "network"
      ]
    }