National Cyber Warfare Foundation (NCWF)login

knock-knock for visualizing live internet break-in attempts against your own honeypots


0 user ratings
2026-11-12 13:32:05
milo
Red Team (CNA)
"knock-knock

knock-knock is a self-hosted, multi-protocol honeypot with a live animated dashboard that captures and visualizes unauthorized scanning and credential-guessing traffic for threat-intelligence research.








Tooldjkurlander/knock-knock — multi-protocol honeypot with live animated attack dashboard and public IP blocklist feed
CategoryDeception technology / honeypot and threat-intelligence visualization
Primary UseDeploying instrumented SSH, Telnet, FTP, RDP, SMB, SIP, HTTP, and SMTP listeners to observe and analyze unsolicited attack traffic on infrastructure you own
Safe UseIntended for authorized security research, lab environments, and defensive threat-intelligence collection on honeypot servers you control; the README notes the bots "can't come in" because the services are simulated, not real
Telemetry NoteAs a honeypot, it is itself a detection instrument: every knock is logged with GeoIP, ASN, ISP, username, password, and protocol, and is republished as an hourly public blocklist consumable by CSF, CrowdSec, ipset, nftables, and pfSense

Every internet-facing server is being probed constantly, and djkurlander/knock-knock turns that ambient hostility into something you can actually watch. The tool is a self-hosted, multi-protocol honeypot paired with a live, animated dashboard that streams every unauthorized connection attempt — a "knock" in the project's vocabulary — in real time. It is written in Python, MIT-licensed, and has accumulated around 207 stars on GitHub, which is respectable traction for a niche defensive tooling project. The framing throughout the README is unapologetically entertaining: you watch bots fail, scoff at offending ISPs, and even hear a virtual geiger counter ticking off what the author calls the background radiation of the Internet.


Under the hood, the coverage is broader than the marketing suggests. Eight core protocols ship out of the box: SSH, Telnet, FTP, RDP, SMB, SIP, HTTP, and SMTP. Beyond those, MQTT, Node-RED, Modbus, S7, and SNMP modules are already included — a notable addition because industrial protocols like Modbus and S7 are rarely first-class citizens in hobbyist honeypots. The README stresses extensibility: each protocol is a self-contained module, so writing and plugging in a new honeypot is presented as a first-class workflow rather than a fork-and-patch exercise. For a research team tracking scanning behavior against OT-adjacent services, that module design matters more than the default protocol list.


The architecture is documented compactly in what the author calls an "Architecture (In One Breath)" diagram, and it is worth unpacking. Traffic flows from the attacker into the protocol-specific honeypots, then into monitor.py, which performs GeoIP enrichment — city, country, ISP, and ASN per attacking IP. Events land in SQLite for persistence while Redis pub/sub handles fan-out, and main.py exposes everything through a FastAPI backend with WebSocket streaming into the live web dashboard. That split is sensible for the use case: SQLite keeps the operational footprint trivial for a single-box deployment, while Redis pub/sub gives you the realtime multiplexing the animated frontend demands.


A second architectural detail deserves attention: the tool supports forwarding knocks from multiple honeypot servers to a single aggregating dashboard. This is where knock-knock graduates from a curiosity into a genuine distributed sensing platform. You can plant lightweight honeypot instances across several networks or cloud regions, stream their events centrally, and the stats panels will break down knock frequency per feeder server. For a blue team building geographic visibility into scanning patterns, that aggregation model — without needing a heavyweight SIEM pipeline — is the practical selling point.


The dashboard itself is the showpiece. A realtime feed shows each knock with location, username, password, ISP, IP, and protocol. Leaderboards track the top offending countries, the most-attempted usernames and passwords, the worst ISPs (the "Wall of Shame"), and the highest-volume IP addresses. A 3D globe visualization — available in six styles, including an extruded country heat map of worst offenders — plots the source of the most recent knock. There is also a Trivia panel that explains why a given username or password may have been chosen, which is a genuinely useful educational touch: it contextualizes credential lists as the recycled residue of prior breaches rather than random noise. The UI works on both desktop and mobile via a swipeable carousel, which suggests the author expects people to leave it running on a screen somewhere.


The most operationally valuable feature, however, is the threat-intelligence output. knock-knock maintains a public blocklist feed of attacker IPs, regenerated hourly, with ready-to-use consumption recipes for CSF, CrowdSec, ipset, nftables, and pfSense. This closes the loop between observation and defense: the honeypot does not just show you who is knocking, it produces an actionable artifact you can feed into firewall policy on your own perimeter. Defenders should treat any such feed with normal caution — honeypot IPs can include shared infrastructure and transient cloud hosts — but as a supplementary signal layered with reputation sources, it is exactly the kind of output a honeypot should generate.


Deployment is deliberately low-friction. The README points to INSTALL.md for full detail and lists three supported paths: Docker (described as simplest and universal), native Ubuntu/Debian, and native RHEL/CentOS/Fedora. The presence of a Docker path plus two distro families without containers means the tool is targeting both the hobbyist running it on a VPS and the operator with deployment constraints. A representative install would be a git clone of the repository followed by the instructions in INSTALL.md, or a docker pull of the packaged image.


There is a public live demo at knock-knock[.]net, along with a summary tutorial page, which is useful for evaluating the dashboard before committing a server to it. Note the demo is defanged here per editorial practice; the README links it directly. The demo also serves a documentary purpose — it is a window into what unsolicited internet traffic looks like at any given moment, which makes it a decent teaching aid for security awareness discussions with non-technical audiences. No victim data is involved; the observed traffic is attackers probing deliberately exposed honeypots.


From a defender's perspective, the telemetry this tool produces is rich and cheap to acquire. Every knock is enriched with GeoIP and ASN context, credential guesses are logged verbatim, and protocol frequency statistics emerge over time. That data supports several legitimate analyses: tracking shifts in brute-force credential preferences, identifying which of your protocols attract the most attention, measuring the dwell-time of recurring scanner IPs, and benchmarking scan volume across regions. Because the events stream over WebSockets, integrating them into your own tooling — an alerting hook, a SIEM ingest, a custom analytics job against the SQLite store — is straightforward.


It is worth being precise about the safety model, because honeypots sit in a sensitive category. knock-knock simulates services rather than exposing real ones — the README's own reassurance is that the bots may be knocking but can't come in. The correct deployment is on infrastructure you own or are authorized to instrument, as a research or early-warning sensor, never as an attempt to retaliate against or deanonymize the actors on the other end. The blocklist feed should be consumed defensively on your own perimeter. Used that way, the tool is squarely educational and documentary in character.


Caveats worth flagging before production use: the README does not discuss authentication on the dashboard itself, so if you self-host the aggregator you should verify how the web UI is exposed and consider putting it behind a VPN or reverse-proxy auth layer, especially since it displays live attacker data. The SQLite + Redis stack implies some operational attention to memory if knock volume is high across many feeders. And as with any honeypot, check the legal posture in your jurisdiction regarding collecting and republishing third-party IP data — the public feed makes that question concrete rather than theoretical.


Taken together, knock-knock is one of the more polished entries in the hobbyist-to-professional honeypot spectrum. It combines credible protocol breadth (including OT protocols), a clean event pipeline through Redis pub/sub and WebSockets, multi-server aggregation, and — critically — an hourly blocklist that turns observation into enforceable policy. For authorized defenders, researchers, and anyone who wants a visceral, data-backed picture of internet background radiation, it earns its place on a sensor VPS.



Official project repository for djkurlander/knock-knock.

Download Tool

Educational analysis for authorized security professionals. Use only in controlled, authorized environments.






Source: OffensiveSec
Source Link: https://www.offsecblog.com/2026/11/knock-knock-for-visualizing-live.html


Comments
new comment
Nobody has commented yet. Will you be the first?
 
Forum
Red Team (CNA)



Copyright 2012 through 2026 - National Cyber Warfare Foundation - All rights reserved worldwide.