Zero-Overhead eBPF Probe Engine v4.18

Kernel-Level Network Telemetry & BGP Latency Analytics

Stream sub-millisecond packet introspection, eBPF kprobe kernel metrics, BGP route flap alerts, and TLS 1.3 handshake benchmarks directly into your edge infrastructure.

Explore API Reference eBPF Kernel Engine

PROBE OVERHEAD < 3.2 nanoseconds

GLOBAL EDGE POPs 142 PoPs / 48 Countries

BGP MONITORING Full ASN Route Table

telemetry-daemon --stream --node=fra-edge-01

LIVE gRPC RAW

WebSocket TLS 1.3 Active

Telemetry Heatmap

Real-Time Global Latency Matrix (p50 / p95 / p99)

Updated: Realtime

p50 Latency 4.12 ms AWS us-east-1 ↔ Frankfurt

p95 Latency 18.40 ms Trans-Atlantic Fiber RTT

p99 Tail Latency 34.18 ms BGP Multi-Hop Route Flaps

Packet Jitter 0.41 ms Kernel Socket Buffer Delta

p50 (Median RTT) p95 (Edge SLA) p99 (Kernel Anomaly)

Kernel Probe Sampling Rate: 1,000 Hz

Kernel & Infrastructure Architecture

Engineered for Low-Latency Cloud Platforms

Standard ICMP pings fail to capture TLS handshake degradation, gRPC frame queueing, or eBPF socket buffer drops. NETSTATS operates directly in kernel space.

eBPF Kernel Probes

Attach non-intrusive kprobe and tracepoint hooks directly to Linux socket syscalls (tcp_v4_connect, tcp_rcv_established) with zero context-switch overhead.

✓ Linux Kernel 5.15+ CO-RE
✓ Socket Buffer Inspection
✓ Zero User-Space Copy

XDP Fast-Path Filter

Parse incoming network frame headers at the eXpress Data Path (XDP) network driver layer before memory allocation by kernel sk_buff.

✓ 100Gbps Line Rate Processing
✓ SYN Flood & UDP Mitigation
✓ Hardware Offload Compatible

BGP Route Flap Alerts

Monitor global AS-path announcements, withdrawal surges, and peer flapping across Tier-1 transit providers (Telia, NTT, Cogent, Lumen) in real-time.

✓ Autonomous System Telemetry
✓ Sub-Second Route Flap Alerts
✓ Prefix Hijack Anomalies

TLS 1.3 Introspection

Dissect client and server TLS handshakes. Measure exact timing for ServerHello, ALPN negotiation, session ticket validation, and certificate chain verification.

✓ Zero-RTT Handshake Metrics
✓ Cipher Suite Performance
✓ mTLS Overhead Breakdown

Edge Presence

142 Telemetry Collectors Worldwide

Deployed directly inside Equinix, AWS, GCP, and Azure Internet Exchange Nodes.

View full edge node list & status →

FRA-01

Frankfurt, DE 1.8 ms

TYO-02

Tokyo, JP 4.1 ms

IAD-04

Ashburn, US 2.4 ms

SIN-01

Singapore, SG 3.9 ms

LHR-03

London, UK 2.1 ms

GRU-01

Sao Paulo, BR 8.7 ms

SYD-02

Sydney, AU 6.2 ms

ZRH-01

Zurich, CH 1.5 ms

Native Integrations

Deploy Telemetry Collectors in Minutes

Integrate NETSTATS telemetry into your Go backend, Rust microservices, Python infrastructure pipelines, or Kubernetes clusters via our lightweight daemonset helm chart.

1 gRPC streaming API with mTLS authentication

2 Native Prometheus & OpenTelemetry exporters

3 Automated BGP route flap webhook dispatch

package main

import (
    "context"
    "log"
    "time"

    "api-netstats.space/sdk/telemetry"
)

func main() {
    client, err := telemetry.NewClient(telemetry.Config{
        APIToken:    "ns_live_9f8a371c20e94b0d88f12a3b",
        Endpoint:    "grpc.api-netstats.space:443",
        EnableeBPF:  true,
        SamplingHz:  1000,
    })
    if err != nil {
        log.Fatalf("Failed to initialize telemetry agent: %v", err)
    }
    defer client.Close()

    // Stream latency percentiles and BGP route flap alerts
    sub := client.SubscribeLatency(context.Background(), "fra-01", "iad-04")
    for event := range sub.C() {
        log.Printf("[METRIC] p99 RTT: %.2fms | BGP AS-Path: %v | Jitter: %.3fms",
            event.P99LatencyMS, event.ASPath, event.JitterMS)
    }
}