Segment Routing Metric Collection

Segment Routing Metrics

Segment Routing SR-MPLS SRv6 Metric Collection Telemetry Prometheus Grafana Monitoring Network Production

Telemetry Pipeline

# === Telemetry Collection Pipeline ===

# Architecture:
# Router (gNMI) → Telegraf/Pipeline → Kafka → Prometheus → Grafana
#
# Cisco IOS-XR Telemetry Config:
# telemetry model-driven
#  destination-group PROMETHEUS
#   address-family ipv4 10.0.0.100 port 57000
#    encoding self-describing-gpb
#    protocol grpc no-tls
#  sensor-group INTERFACE
#   sensor-path Cisco-IOS-XR-infra-statsd-oper:infra-statistics/interfaces/interface/latest/generic-counters
#  sensor-group SR-POLICY
#   sensor-path Cisco-IOS-XR-infra-xtc-agent-oper:xtc/policy-forwardings
#  sensor-group IGP
#   sensor-path Cisco-IOS-XR-clns-isis-oper:isis/instances/instance/neighbors
#  subscription MONITORING
#   sensor-group-id INTERFACE sample-interval 10000
#   sensor-group-id SR-POLICY sample-interval 30000
#   sensor-group-id IGP sample-interval 60000
#   destination-id PROMETHEUS

from dataclasses import dataclass

@dataclass
class TelemetrySource:
    source: str
    protocol: str
    yang_model: str
    interval: str
    data_points: str

sources = [
    TelemetrySource("Interface Counters",
        "gNMI / MDT (Model-Driven Telemetry)",
        "Cisco-IOS-XR-infra-statsd-oper",
        "10 วินาที",
        "bytes_in/out, packets_in/out, errors, drops"),
    TelemetrySource("SR Policy Status",
        "gNMI / MDT",
        "Cisco-IOS-XR-infra-xtc-agent-oper",
        "30 วินาที",
        "policy_name, color, endpoint, active_path, segment_list"),
    TelemetrySource("ISIS Neighbors",
        "gNMI / MDT",
        "Cisco-IOS-XR-clns-isis-oper",
        "60 วินาที (event-driven)",
        "neighbor_id, state, interface, level, hold_time"),
    TelemetrySource("Performance Measurement",
        "gNMI / MDT",
        "Cisco-IOS-XR-perf-meas-oper",
        "10 วินาที",
        "delay_min/max/avg per link, jitter, loss"),
    TelemetrySource("Traffic Flow (sFlow)",
        "sFlow v5 / NetFlow v9",
        "N/A (packet sampling)",
        "1:1000 sampling",
        "src/dst IP, port, protocol, bytes, application"),
]

print("=== Telemetry Sources ===")
for s in sources:
    print(f"  [{s.source}] Protocol: {s.protocol}")
    print(f"    YANG: {s.yang_model}")
    print(f"    Interval: {s.interval}")
    print(f"    Data: {s.data_points}")

Prometheus + Grafana

# === Monitoring Stack ===

# Telegraf config for gNMI input
# [[inputs.gnmi]]
#   addresses = ["router1:57400", "router2:57400"]
#   username = "admin"
#   password = "secret"
#   [[inputs.gnmi.subscription]]
#     name = "interface"
#     origin = "Cisco-IOS-XR-infra-statsd-oper"
#     path = "/infra-statistics/interfaces/interface/latest/generic-counters"
#     subscription_mode = "sample"
#     sample_interval = "10s"
#   [[inputs.gnmi.subscription]]
#     name = "sr_policy"
#     origin = "Cisco-IOS-XR-infra-xtc-agent-oper"
#     path = "/xtc/policy-forwardings"
#     subscription_mode = "sample"
#     sample_interval = "30s"
#
# [[outputs.prometheus_client]]
#   listen = ":9273"
#   metric_version = 2

# Prometheus scrape config
# scrape_configs:
#   - job_name: 'telegraf'
#     static_configs:
#       - targets: ['telegraf:9273']
#   - job_name: 'snmp'
#     static_configs:
#       - targets: ['snmp-exporter:9116']

# Grafana Alert Rules (example)
# - alert: HighLinkUtilization
#   expr: rate(interface_bytes_total[5m]) * 8 / interface_speed > 0.8
#   for: 5m
#   labels:
#     severity: warning
#   annotations:
#     summary: "Link {{ $labels.interface }} utilization > 80%"

@dataclass
class GrafanaPanel:
    panel: str
    query: str
    visualization: str
    alert: str

panels = [
    GrafanaPanel("Link Utilization (%)",
        "rate(interface_bytes_total[5m]) * 8 / interface_speed * 100",
        "Time Series + Threshold lines at 70% 80%",
        "> 80% for 5m → Warning, > 90% → Critical"),
    GrafanaPanel("SR Policy Status",
        "sr_policy_state{state='active'} == 1",
        "Stat Panel: Green=Active, Red=Down",
        "state != active → Critical Alert"),
    GrafanaPanel("Per-Link Latency",
        "perf_meas_delay_avg{direction='forward'}",
        "Heatmap: Links × Time → Color = Delay",
        "> 10ms → Warning, > 50ms → Critical"),
    GrafanaPanel("Top 10 Utilized Interfaces",
        "topk(10, rate(interface_bytes_total[5m]) * 8)",
        "Bar Chart: sorted by utilization",
        "N/A (informational)"),
    GrafanaPanel("IGP Topology Changes",
        "increase(isis_adjacency_changes_total[1h])",
        "Time Series: spikes = topology flaps",
        "> 5 changes/hr → Warning"),
]

print("=== Grafana Panels ===")
for p in panels:
    print(f"  [{p.panel}]")
    print(f"    Query: {p.query}")
    print(f"    Viz: {p.visualization}")
    print(f"    Alert: {p.alert}")

Automation

# === Network Automation for SR ===

# Python script to collect SR policy info via SSH
# from netmiko import ConnectHandler
#
# device = {
#     'device_type': 'cisco_xr',
#     'host': '10.0.0.1',
#     'username': 'admin',
#     'password': 'secret',
# }
#
# conn = ConnectHandler(**device)
# output = conn.send_command('show segment-routing traffic-eng policy color 100')
# sr_policies = conn.send_command('show segment-routing traffic-eng policy all')
# perf_data = conn.send_command('show performance-measurement interfaces')
# conn.disconnect()

@dataclass
class AutomationTask:
    task: str
    tool: str
    trigger: str
    action: str

tasks = [
    AutomationTask("Link Down Response",
        "Ansible + Event-driven",
        "IGP Adjacency Down event",
        "Verify SR Policy reroute สำเร็จ แจ้ง NOC ถ้าไม่สำเร็จ"),
    AutomationTask("Capacity Planning",
        "Python + Prometheus API",
        "Weekly scheduled job",
        "ดึง Utilization data วิเคราะห์ Trend แจ้งเมื่อใกล้ Capacity"),
    AutomationTask("SR Policy Validation",
        "Python + gNMI",
        "หลัง Config Change",
        "ตรวจว่า SR Policy Active ถูกต้อง Latency ไม่เปลี่ยน"),
    AutomationTask("Topology Backup",
        "Python + Netconf",
        "ทุกวัน 02:00",
        "Backup Configuration ทุก Router เก็บ Git"),
    AutomationTask("SLA Report",
        "Python + Grafana API",
        "ทุกเดือน วันที่ 1",
        "สร้าง PDF Report แสดง Uptime Latency Utilization"),
]

print("=== Automation Tasks ===")
for t in tasks:
    print(f"  [{t.task}] Tool: {t.tool}")
    print(f"    Trigger: {t.trigger}")
    print(f"    Action: {t.action}")

เคล็ดลับ

• Telemetry: ใช้ Streaming Telemetry แทน SNMP ข้อมูลละเอียดกว่า
• gNMI: ใช้ gNMI สำหรับ Router รุ่นใหม่ ประสิทธิภาพสูงกว่า
• Kafka: ใช้ Kafka เป็น Buffer เมื่อมี Router หลายพันตัว
• Alert: ตั้ง Alert Link Utilization > 80% Latency > Threshold
• Automation: ใช้ Ansible Python ทำ Validation หลัง Config Change

Segment Routing คืออะไร

SR-MPLS SRv6 Segment Label SID Routing Architecture IGP OSPF IS-IS Traffic Engineering Fast Reroute SR-PCE Controller ISP Data Center WAN

สรุป

Segment Routing SR-MPLS SRv6 Telemetry gNMI Prometheus Grafana Monitoring Interface Latency SR Policy IGP Automation Production