Skip to content

DDoS & Dstat Glossary

Plain-language definitions of the terms you will see on a Dstat dashboard — from Layer 4 vs Layer 7 to RPS, Under Attack Mode and bypass — with notes on how each one shows up on the live chart.

Dstat

Live traffic statistics during a stress test or DDoS attack.

Dstat (short for "destination statistics") is a live view of the request/bandwidth volume a target receives during a stress test, typically a rolling chart of requests per time window plus a breakdown of passed, bypassed and blocked traffic. Multiple such graphs are called Dstats. On this dashboard the chart plots requests over a rolling 180-second window, so a sustained climb means the target is absorbing more load than its baseline.

Layer 7

Application-layer traffic measured in requests per second.

Layer 7 (the application layer) covers HTTP/HTTPS requests. Layer7 Dstat is measured in requests, and is where WAFs, JS challenges, CAPTCHAs and rate limits do their work. The headline number is peak requests per second (RPS), and a protected target should show most of that volume filtered before it reaches the origin.

Layer 4

Network/transport-layer traffic measured in bandwidth and packets.

Layer 4 (the transport layer) covers TCP/UDP volume. Layer4 Dstat is measured in bandwidth (Gbps) and packets per second (pps), independent of any application logic. Because these floods ignore the application, they are mitigated at the network edge rather than by challenges or rate limits.

Under Attack Mode (UAM)

A JavaScript challenge that filters automated clients.

Under Attack Mode issues a JavaScript challenge that a real browser solves transparently while most simple bots fail, sharply reducing automated Layer-7 traffic that reaches the origin. On a Dstat chart, enabling UAM typically shows up as a sharp drop in bypass traffic shortly after the test begins.

CAPTCHA

An interactive challenge requiring human input.

A CAPTCHA challenge requires explicit human interaction (clicking or solving a puzzle). It is stronger than a silent JS challenge but adds friction for legitimate users, so it is usually reserved for high-risk paths such as login or checkout where a drop in bypass traffic is worth the extra friction.

Rate Limiting

Capping requests per client over a time window.

Rate limiting caps how many requests a single client (IP, token, or path) may send within a window, smoothing bursts and blunting volumetric Layer-7 floods. In Dstat terms, effective rate limiting keeps the bypass line flat even as total incoming requests spike.

RPS (Requests Per Second)

The core Layer-7 throughput metric.

Requests per second is the primary unit for Layer7 Dstat: it expresses how many HTTP requests a target processes each second, and peak RPS is the headline figure of most stress tests. Comparing incoming RPS against bypass RPS on the same chart is a quick way to read how well a protection is holding.

Bypass

Requests that slipped past protections to the origin.

Bypass counts requests that defeated the protection layer (challenge or filter) and reached the origin. A high bypass rate during a test indicates the protection is not fully effective. The goal of any protected Dstat run is to drive the bypass line toward zero while total request volume stays high.

pps (Packets Per Second)

The core Layer-4 packet-rate metric.

Packets per second counts how many network packets reach a target each second, regardless of their size. It is the second headline figure of a Layer4 Dstat next to bandwidth, and it matters because routers, firewalls and network cards have a packet-handling limit of their own: a flood of very small packets can exhaust that limit while the bandwidth graph still looks modest.

Gbps (Gigabits Per Second)

The bandwidth unit of a Layer4 Dstat.

Gbps is one billion bits per second and is the unit a Layer4 Dstat uses for bandwidth. Network capacity is quoted in bits while file sizes are quoted in bytes, so 8 Gbps equals about 1 gigabyte per second of traffic. Larger floods are written in Tbps, where 1 Tbps equals 1,000 Gbps.

UDP Flood

A volumetric flood of UDP packets.

A UDP flood sends a large number of UDP packets to random or fixed ports on a target. UDP has no handshake, so the sender needs no reply and the target has to spend resources receiving and discarding every packet. On a Layer4 Dstat it appears as a steep rise in both Gbps and pps, and it is the typical shape of a pure capacity attack.

SYN Flood

A TCP flood that exhausts connection state.

A SYN flood sends TCP connection requests (SYN packets) without completing the handshake, leaving the target with half-open connections that occupy its connection table. Each packet is tiny, so a SYN flood tends to show a high pps with a modest Gbps figure. This is the main reason a Layer4 Dstat tracks packets as well as bandwidth.

Amplification Attack

A reflection attack that multiplies traffic through third-party servers.

In an amplification attack the sender spoofs the victim's address and queries a public service, such as DNS or NTP, that answers with a much larger response. The victim receives the amplified replies from many servers it never contacted. On a Layer4 Dstat this produces very high Gbps from a small amount of attacker traffic, which is why it is mitigated upstream with scrubbing rather than at the application.

Learn more