Data Collection
Infrastructure Metrics
We collect the following metrics from our network infrastructure:
Router Metrics (via MKTXP)
| Metric Category | Data Points | Retention |
|---|---|---|
| Interface Statistics | TX/RX bytes, packets, errors, drops | 30 days |
| WireGuard Peers | Handshake time, transfer stats per peer | 30 days |
| System Resources | CPU, memory, disk, temperature | 30 days |
| Firewall Counters | Rule hit counts, connection tracking | 30 days |
| Routing Tables | BGP/OSPF state, route counts | 30 days |
| IP Connections | Active connection count, protocol breakdown | 30 days |
Host Metrics (via Node Exporter)
| Metric Category | Data Points |
|---|---|
| CPU | Utilisation, load average, per-core stats |
| Memory | Used, available, cached, swap |
| Disk | I/O operations, latency, space utilisation |
| Network | Interface throughput, errors, TCP connections |
| System | Uptime, kernel version, process counts |
Container Metrics (via cAdvisor)
| Metric Category | Data Points |
|---|---|
| Resource Usage | CPU, memory per container |
| Network I/O | Container network throughput |
| Filesystem | Container storage usage |
Traffic Flow Analysis
NetFlow v9 data is exported from our core router and collected by pmacct for deep traffic analysis:
| Analysis Type | Description |
|---|---|
| Protocol Distribution | Breakdown of traffic by IP protocol (TCP, UDP, ICMP, etc.) |
| Top Talkers | Highest volume source/destination IP pairs |
| Port Analysis | Traffic distribution by source and destination ports |
| Bandwidth Trends | Historical traffic patterns over time |