Network & IoT Integrations

This document covers network device and IoT messaging integrations.

Overview

flowchart TB
    subgraph "TelemetryFlow Agent"
        IM[Integration Manager]
    end

    subgraph "Network Management"
        DNAC[Cisco DNA Center]
        MERAKI[Cisco Meraki]
    end

    subgraph "Network Protocols"
        SNMP[SNMP v1/v2c/v3]
    end

    subgraph "IoT Messaging"
        MQTT[MQTT Broker]
    end

    IM --> DNAC & MERAKI
    IM --> SNMP
    IM --> MQTT

    DNAC --> |Devices, Health| IM
    MERAKI --> |Organizations, Devices| IM
    SNMP --> |OIDs, MIBs| IM
    MQTT --> |Topics| IM

Cisco (DNA Center / Meraki)

Architecture

sequenceDiagram
    participant Agent as TFO Agent
    participant DNAC as DNA Center
    participant Meraki as Meraki Dashboard

    alt DNA Center
        Agent->>DNAC: POST /dna/system/api/v1/auth/token
        DNAC-->>Agent: Auth Token

        Agent->>DNAC: GET /dna/intent/api/v1/network-device
        DNAC-->>Agent: Device List

        Agent->>DNAC: GET /dna/intent/api/v1/network-health
        DNAC-->>Agent: Health Scores
    else Meraki
        Agent->>Meraki: GET /organizations (X-Cisco-Meraki-API-Key)
        Meraki-->>Agent: Organizations

        Agent->>Meraki: GET /organizations/{id}/devices/statuses
        Meraki-->>Agent: Device Statuses
    end

Configuration

integrations:
  cisco:
    enabled: true
    api_type: dnac # or meraki

    # DNA Center configuration
    dnac:
      endpoint: "https://dnac.example.com"
      username: "${CISCO_DNAC_USERNAME}"
      password: "${CISCO_DNAC_PASSWORD}"

    # Meraki configuration
    meraki:
      api_key: "${CISCO_MERAKI_API_KEY}"
      # org_id: "${CISCO_MERAKI_ORG_ID}"

    tls_skip_verify: false
    scrape_interval: 60s
    timeout: 30s

    collect_devices: true
    collect_networks: true
    collect_clients: true
    collect_health: true
    collect_events: false

Metrics

DNA Center

Metric	Type	Description
`cisco_dnac_device_up`	gauge	Device reachability (1=up)
`cisco_dnac_device_uptime_seconds`	gauge	Device uptime
`cisco_dnac_network_health_score`	gauge	Network health (0-100)
`cisco_dnac_client_health_score`	gauge	Client health (0-100)
`cisco_dnac_application_health_score`	gauge	App health (0-100)

Meraki

Metric	Type	Description
`cisco_meraki_device_online`	gauge	Device online status
`cisco_meraki_device_using_cellular`	gauge	Cellular failover active
`cisco_meraki_organization_devices_total`	gauge	Total devices

SNMP

Architecture

sequenceDiagram
    participant Agent as TFO Agent
    participant Device as Network Device

    alt SNMPv2c
        Agent->>Device: GET (community string)
        Device-->>Agent: OID Values
    else SNMPv3
        Agent->>Device: GET (user/auth/priv)
        Device-->>Agent: OID Values
    end

    loop For each OID
        Agent->>Device: SNMP GET/WALK
        Device-->>Agent: Value
    end

Configuration

integrations:
  snmp:
    enabled: true
    version: v2c # v1, v2c, v3
    community: "${SNMP_COMMUNITY:-public}"
    port: 161
    timeout: 10s
    retries: 3
    scrape_interval: 60s
    max_repetitions: 10

    # SNMPv3 settings
    security_level: noAuthNoPriv # noAuthNoPriv, authNoPriv, authPriv
    username: "${SNMP_USERNAME}"
    auth_protocol: SHA # MD5, SHA
    auth_password: "${SNMP_AUTH_PASSWORD}"
    priv_protocol: AES # DES, AES
    priv_password: "${SNMP_PRIV_PASSWORD}"

    targets:
      - address: "switch1.example.com"
        port: 161
        name: "Core Switch 1"
        community: "private"
        labels:
          location: datacenter1

      - address: "router1.example.com"
        name: "Edge Router 1"

    # walk_oids / get_oids are optional. When omitted, the agent applies the
    # industry-standard defaults below (system scalars + full IF-MIB tables).
    walk_oids:
      - "1.3.6.1.2.1.2.2" # IF-MIB::ifTable
      - "1.3.6.1.2.1.31.1.1.1" # IF-MIB::ifXTable (64-bit HC counters, ifName)

    get_oids:
      - oid: "1.3.6.1.2.1.1.3.0"
        name: "sysUpTime"
        type: counter
      - oid: "1.3.6.1.2.1.1.5.0"
        name: "sysName"
        type: string

Default OIDs

When no get_oids / walk_oids are configured, the agent polls a standard baseline. Scalars are collected via SNMP GET; per-interface metrics are collected by WALKing the IF-MIB tables, so a fresh target yields useful interface data with no manual OID mapping.

Scalar GET defaults (SNMPv2-MIB / UCD-SNMP-MIB):

OID	Name	Type	Unit
1.3.6.1.2.1.1.1.0	sysDescr	string	—
1.3.6.1.2.1.1.3.0	sysUpTime	counter	ticks
1.3.6.1.2.1.1.5.0	sysName	string	—
1.3.6.1.4.1.2021.11.9.0	ssCpuUser	gauge	percent
1.3.6.1.4.1.2021.11.11.0	ssCpuIdle	gauge	percent
1.3.6.1.4.1.2021.4.5.0	memTotalReal	gauge	kB
1.3.6.1.4.1.2021.4.6.0	memAvailReal	gauge	kB

Interface WALK defaults (IF-MIB — RFC 1213 + RFC 2233):

Table root	Provides
1.3.6.1.2.1.2.2	`ifTable` — ifSpeed, ifOperStatus, ifIn/OutOctets, errors, discards
1.3.6.1.2.1.31.1.1.1	`ifXTable` — ifName, ifHCIn/OutOctets (64-bit), ifHighSpeed

64-bit counters: ifXTable HC counters are collected because 32-bit ifInOctets/ifOutOctets wrap in seconds on ≥1 Gbps links and cannot be used for accurate rate calculation. This matches standard NMS practice (LibreNMS, Observium, Prometheus snmp_exporter).

Each walked row carries an index tag (the interface index) so per-interface series stay distinct — feeding the platform's Interface Utilization view (in/out, capacity, errors, discards, oper status) directly.

Metrics

Metric	Type	Source (IF-MIB)	Description
`snmp_target_up`	gauge	probe	Target reachability (0/1)
`snmp_sysuptime`	counter	sysUpTime	System uptime ticks
`snmp_walk_1_3_6_1_2_1_31_1_1_1_6`	counter	ifHCInOctets	Interface bytes in (64-bit)
`snmp_walk_1_3_6_1_2_1_31_1_1_1_10`	counter	ifHCOutOctets	Interface bytes out (64-bit)
`snmp_walk_1_3_6_1_2_1_2_2_1_14`	counter	ifInErrors	Interface inbound errors
`snmp_walk_1_3_6_1_2_1_2_2_1_13`	counter	ifInDiscards	Interface inbound discards
`snmp_walk_1_3_6_1_2_1_2_2_1_8`	gauge	ifOperStatus	Interface operational status
`snmp_sscpuuser`	gauge	ssCpuUser	CPU user percentage

Walk metric names default to snmp_walk_<oid>; assign a friendly name in a get_oids/MIB mapping to rename them.

Operational Notes

Concurrent polling: targets are polled in parallel (bounded pool) so one slow or timing-out device does not delay the rest of the scrape.
Health check: Health() issues a real SNMP GET for sysUpTime.0 per target — a device counts as reachable only if it actually answers SNMP (a UDP socket opening is not treated as reachability).
SNMPv3: privacy (encryption) is enabled whenever security_level: authPriv is set; auth-only and no-auth levels are honored independently.
Cancellation: in-flight GET/WALK I/O is bounded by timeout and unwound on context cancellation (agent shutdown) so no poll leaks a connection.

MQTT

Architecture

sequenceDiagram
    participant Agent as TFO Agent
    participant Broker as MQTT Broker
    participant Sub as Subscribers

    Agent->>Broker: CONNECT (ClientID, Credentials)
    Broker-->>Agent: CONNACK

    Note over Agent,Broker: Last Will configured

    loop Publishing
        Agent->>Broker: PUBLISH (metrics/logs/traces)
        Broker-->>Agent: PUBACK (QoS 1/2)
        Broker->>Sub: Forward Message
    end

    Agent->>Broker: DISCONNECT

Configuration

integrations:
  mqtt:
    enabled: true
    broker: "tcp://mqtt.example.com:1883"
    # broker: "ssl://mqtt.example.com:8883"
    # broker: "ws://mqtt.example.com:8083/mqtt"
    client_id: "${HOSTNAME}-tfo-agent"
    username: "${MQTT_USERNAME}"
    password: "${MQTT_PASSWORD}"

    metrics_topic: "telemetryflow/metrics"
    logs_topic: "telemetryflow/logs"
    traces_topic: "telemetryflow/traces"
    topic_prefix: ""

    qos: 1 # 0, 1, or 2
    retained: false
    clean_session: true

    connect_timeout: 30s
    keep_alive: 60s
    ping_timeout: 10s
    auto_reconnect: true
    max_reconnect_interval: 5m

    tls_enabled: false
    tls_ca_file: ""
    tls_cert_file: ""
    tls_key_file: ""
    tls_skip_verify: false

    encoding: json # json, protobuf
    batch_size: 100
    flush_interval: 10s

    will:
      enabled: true
      topic: "telemetryflow/status/${HOSTNAME}"
      payload: '{"status": "offline"}'
      qos: 1
      retained: true

Topic Structure

telemetryflow/
├── metrics              # Metric data
├── logs                 # Log entries
├── traces               # Trace spans
└── status/
    └── {hostname}       # Agent status (LWT)

Message Format (JSON)

Metrics:

{
  "timestamp": "2024-12-29T10:30:00Z",
  "name": "cpu_usage",
  "value": 45.5,
  "type": "gauge",
  "tags": {
    "host": "server1"
  }
}

Logs:

{
  "timestamp": "2024-12-29T10:30:00Z",
  "level": "info",
  "message": "Service started",
  "source": "app.main",
  "trace_id": "abc123"
}

QoS Levels

QoS	Name	Guarantee
0	At most once	Fire and forget
1	At least once	Acknowledged delivery
2	Exactly once	Two-phase commit

Comparison

Feature	Cisco DNAC	Cisco Meraki	SNMP	MQTT
Protocol	REST API	REST API	UDP	TCP
Auth	Token	API Key	Community/USM	User/Pass
Direction	Pull	Pull	Pull	Push
Metrics	✅	✅	✅	✅
Logs	❌	❌	❌	✅
Traces	❌	❌	❌	✅
Real-time	❌	❌	❌	✅

Navigation Menu

Uh oh!

Network & IoT Integrations

Overview

Cisco (DNA Center / Meraki)

Architecture

Configuration

Metrics

DNA Center

Meraki

SNMP

Architecture

Configuration

Default OIDs

Metrics

Operational Notes

MQTT

Architecture

Configuration

Topic Structure

Message Format (JSON)

QoS Levels

Comparison

Uh oh!

FilesExpand file tree

NETWORK.md

Latest commit

History

NETWORK.md

File metadata and controls

Network & IoT Integrations

Overview

Cisco (DNA Center / Meraki)

Architecture

Configuration

Metrics

DNA Center

Meraki

SNMP

Architecture

Configuration

Default OIDs

Metrics

Operational Notes

MQTT

Architecture

Configuration

Topic Structure

Message Format (JSON)

QoS Levels

Comparison