Overview

Nekotopia operates a comprehensive monitoring infrastructure to ensure service reliability, performance optimisation, and proactive issue detection. Our telemetry stack is built on industry-standard open-source tools,tools.

providing

---

real-time

What visibilityWe intoMonitor

network

We continuously monitor the health and VPNperformance serviceof performance.the Torus network:

---

Monitoring Stack

Core Components

WhattunnelCorecapacity

ComponentArea	Purpose	DescriptionWe Track
PrometheusNetwork Health	MetricsVPN Database	Time-seriesstatus, collectionbandwidth andutilisation, storagepacket with powerful query language (PromQL)loss
GrafanaService Availability	Visualisation	Dashboards,services alerting,uptime, andAPI data explorationresponsiveness
MKTXPResource Usage	MikroTikInfrastructure Exporter	Exportsand RouterOS metrics (interfaces, queues, firewall, wireless)headroom
NodeSecurity ExporterEvents	Host Metrics	CPU, memory, disk, network stats from Linux servers
cAdvisor	Container Metrics	Docker container resource usage and performance
pmacct	Flow Analysis	NetFlow/IPFIXUnusual traffic accountingpatterns, andconnection analysisanomalies

---

ArchitectureAlerting

TheAutomated monitoringalerts systemnotify followsadministrators a pull-based collection model with centralised storage and visualisation.

Data Collection Flow

1. Data Sources generate metrics:of:

• MikroTikService Routerdegradation (networkor stats, queues, firewall)outages
• LinuxCapacity Hoststhresholds (systemapproaching resources)limits
• DockerSecurity-relevant Containersevents
(application
• Infrastructure metrics)component failures

---

Dashboards

Real-time dashboards provide visibility into:

• Overall network health status
• Active VPN connections
• Bandwidth consumption by tier
• Historical performance trends

---

User-Facing Metrics

Some metrics are exposed to users through the dashboard:

• Your connection status and uptime
• Your bandwidth usage
• Network latency to the hub

2.For Exportersmore expose metrics in Prometheus format:

• MKTXP → MikroTik metricsdetails on :9436
• Node Exporter → Host metrics on :9100
• cAdvisor → Container metrics on :8080

3. Prometheus scrapes all exporters every 15 seconds and stores 30 days of time-series data.

4. Grafana queries Prometheus and displays dashboards for operators.

Visual Flow

╔═══════════════════════════════════════════════════════════════════╗
║                        DATA SOURCES                               ║
╠═══════════════════════════════════════════════════════════════════╣
║                                                                   ║
║   ┌─────────────┐    ┌─────────────┐    ┌─────────────┐          ║
║   │  MikroTik   │    │   Linux     │    │   Docker    │          ║
║   │   Router    │    │   Hosts     │    │ Containers  │          ║
║   └──────┬──────┘    └──────┬──────┘    └──────┬──────┘          ║
║          │                  │                  │                 ║
║          ▼                  ▼                  ▼                 ║
║   ┌─────────────┐    ┌─────────────┐    ┌─────────────┐          ║
║   │    MKTXP    │    │    Node     │    │  cAdvisor   │          ║
║   │  :9436      │    │  :9100      │    │  :8080      │          ║
║   └──────┬──────┘    └──────┬──────┘    └──────┬──────┘          ║
╚══════════╪══════════════════╪══════════════════╪══════════════════╝
           │                  │                  │
           └──────────────────┼──────────────────┘
                              │
                              ▼
              ╔═══════════════════════════════╗
              ║        PROMETHEUS             ║
              ║      (Metrics Database)       ║
              ║                               ║
              ║  • Scrapes every 15 sec       ║
              ║  • 30 day retention           ║
              ║  • Alerting rules             ║
              ╚═══════════════╤═══════════════╝
                              │
                              ▼
              ╔═══════════════════════════════╗
              ║          GRAFANA              ║
              ║        (Dashboards)           ║
              ║                               ║
              ║  • VPN Statistics             ║
              ║  • Bandwidth Graphs           ║
              ║  • System Health              ║
              ╚═══════════════════════════════╝

NetFlow Traffic Analysis

For detailed traffic analysis, NetFlowwhat data followswe acollect separateabout path:users, see Data Collection.

┌─────────────┐         ┌─────────────┐         ┌─────────────┐
│  MikroTik   │ NetFlow │   pmacct    │ metrics │ Prometheus  │
│   Router    │ ──────▶ │  Collector  │ ──────▶ │             │
│             │   v9    │             │         │             │
└─────────────┘         └──────┬──────┘         └─────────────┘
                               │
                               ▼
                        ┌─────────────┐
                        │   Grafana   │
                        │  Traffic    │
                        │  Analysis   │
                        └─────────────┘

---

Key Metrics Collected

Network Metrics

Metric	Description
Interface Throughput	Bytes in/out per interface
Packet Rates	Packets per second, errors, drops
Queue Statistics	Per-user bandwidth enforcement
Firewall Counters	Rule hit counts, blocked connections
WireGuard Peers	Handshake status, data transfer

System Metrics

Metric	Description
CPU/Memory	Server and router resource usage
Disk I/O	Storage performance
Container Health	Docker service status
Process Monitoring	Critical service uptime

Traffic Analysis

Metric	Description
Flow Records	Source/destination IP, ports, protocols
User Bandwidth	Per-VPN-peer traffic accounting
Top Talkers	Highest bandwidth consumers
Protocol Distribution	Traffic breakdown by application

---

Access

Grafana dashboards are available to administrators at the internal monitoring endpoint. User-facing statistics are exposed through the Nekotopia dashboard where appropriate.