Add HAProxy PoW challenge, simple bad bot blocking, and adjust mounts
All checks were successful
Build And Deploy / build-and-deploy (push) Successful in 1m16s

Replace single deploy/haproxy.cfg with deploy/haproxy/{haproxy.cfg,challenge.html}.
HAProxy now runs a WebCrypto-based proof-of-work challenge using a stick-table,
URI normalization and a challenge backend. docker-compose mounts the haproxy
directory, and also switches the site DB volume to ./db to be consistent. Update robots.txt.ts to
add a honeypot path for bad bot blocking.
This commit is contained in:
badblocks 2026-02-08 12:50:28 -08:00
parent 1162c53c8f
commit f014330b14
No known key found for this signature in database
5 changed files with 233 additions and 45 deletions

View file

@ -4,7 +4,7 @@ services:
restart: always restart: always
container_name: badblocks-personal-site container_name: badblocks-personal-site
volumes: volumes:
- /srv/badblocks-personal-site/db:/db - ./db:/db
networks: networks:
- proxynet - proxynet
env_file: env_file:
@ -70,7 +70,7 @@ services:
- "${PUBLIC_IP}:443:443" - "${PUBLIC_IP}:443:443"
- "${PUBLIC_IP}:8404:8404" - "${PUBLIC_IP}:8404:8404"
volumes: volumes:
- ./haproxy.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro - ./haproxy:/usr/local/etc/haproxy:ro
- ./certs:/certs:ro - ./certs:/certs:ro
restart: always restart: always
networks: networks:

View file

@ -1,43 +0,0 @@
global
daemon
log stdout format raw local0 info
maxconn 2000
defaults
mode http
log global
timeout connect 5s
timeout client 30s
timeout server 30s
timeout check 5s
retries 3
option httplog
option dontlognull
option redispatch
frontend http
bind :80
mode http
http-request redirect scheme https unless { ssl_fc }
frontend https
bind :443 ssl crt /certs/fullcert.pem
http-response set-header Strict-Transport-Security "max-age=16000000; includeSubDomains; preload;"
default_backend main
backend main
balance leastconn
option httpchk GET /
http-check expect status 200
server badblocks-personal-site badblocks-personal-site:4321 check resolvers docker resolve-prefer ipv4 init-addr none
resolvers docker
nameserver dns1 127.0.0.11:53
resolve_retries 3
timeout resolve 1s
timeout retry 1s
hold valid 10s
hold obsolete 30s

View file

@ -0,0 +1,143 @@
<!doctype html>
<meta name="viewport" content="width=device-width" />
<title>Challenge Accepted!</title>
<style>
body {
background: #ddd;
color: #000;
margin: 0;
padding: 0;
}
#progress {
position: absolute;
margin: 0;
top: 50%%;
left: 50%%;
transform: translate(-50%%, -50%%);
text-align: center;
font-size: 125%%;
}
#progressBar {
font-size: 250%%;
user-select: none;
}
#progressBar.done {
transition:
font-size 2s linear,
opacity 2s linear;
font-size: 1000%%;
opacity: 0.5;
}
.animate {
animation: spin 2s infinite linear;
}
@keyframes spin {
to {
transform: rotate(360deg);
}
}
iframe {
display: none;
}
footer {
position: fixed;
bottom: 4px;
width: 100%%;
text-align: center;
}
</style>
<form method="post" target="post" action="/_challenge" name="challenge">
<input type="hidden" name="ip" value="%[src]" />
<input type="hidden" name="ts" value="%[date]" />
<input type="hidden" name="diff" value="4" />
<input type="hidden" name="tries" value="" />
</form>
<div id="progress">
<span id="progressText"></span>
<div id="progressBar"></div>
</div>
<iframe srcdoc="" src="about:blank" name="post"></iframe>
<script>
"use strict";
async function challenge(diff, ip, ts) {
let te = new TextEncoder();
let tries = 0;
progressText.innerText = "✋ Checking connection, please wait";
progressBar.innerText = "🌀";
progressBar.className = "animate";
for (; tries < 10_000_000; tries++) {
let hash = await crypto.subtle.digest(
"SHA-256",
te.encode([ip, location.hostname, ts, tries].join(";")),
);
let y = new Uint8Array(hash);
let i = 0;
while (i < diff / 2) {
if (y[i] > 0x0f) break;
if (i * 2 + 1 >= diff) return tries;
if (y[i++] > 0) break;
if (i * 2 >= diff) return tries;
}
}
}
async function startChallenge(form) {
if (!("subtle" in window.crypto)) {
if (location.protocol !== "https:") {
progress.innerText =
"No WebCrypto support. This must be served over a HTTPS connection.";
} else {
progress.innerText =
"No WebCrypto support in your browser. " +
"This is required to pass the challenge.";
}
return;
}
let backoff = 0;
let tryOnce = (_) => {
if (backoff > 8) {
progressText.innerText =
"Failed to submit after several tries. Try reloading.";
return;
}
setTimeout(
async (_) => submitAnswer(form),
1000 * (Math.pow(2, backoff++) - 1),
);
};
form.addEventListener("error", tryOnce);
let iframe = document.querySelector("iframe");
iframe.addEventListener("load", (_) =>
location.hash.length ? location.reload() : location.replace(location),
);
tryOnce();
}
async function submitAnswer(form) {
let start = new Date();
let tries = await challenge(form.diff.value, form.ip.value, form.ts.value);
if (tries === undefined) {
progressText.innerText =
"Unable to calculate challenge. Try reloading or a different browser.";
progressBar.innerText = "🤯";
progressBar.className = "error";
return;
}
form.tries.value = tries;
progressText.innerText = "Took " + (new Date() - start) + "ms";
console.log(`${tries} tries.\n${progressText.innerText}`);
progressBar.className = "done";
progressBar.innerText = "✅";
form.submit();
}
window.addEventListener("load", (_) =>
startChallenge(document.forms.challenge),
);
</script>
<noscript
>Malicious scrapers break the web. To continue, you'll need JavaScript
enabled.</noscript
>

View file

@ -0,0 +1,85 @@
global
daemon
log stdout format raw local0 info
maxconn 2000
# For normalize-uri
expose-experimental-directives
defaults
mode http
log global
timeout connect 5s
timeout client 30s
timeout server 30s
timeout check 5s
retries 3
option httplog
option dontlognull
option redispatch
frontend http
bind :80
mode http
http-request redirect scheme https unless { ssl_fc }
frontend www
bind :443 ssl crt /certs/fullcert.pem
# 2 general purpose tags in this stick-table (name defaults to frontend name, i.e. www)
stick-table type ipv6 size 1m expire 2d store gpt(2)
http-request track-sc0 src
http-request normalize-uri path-merge-slashes
http-request normalize-uri path-strip-dot
http-request normalize-uri path-strip-dotdot
# Drop the connection immediately if the requester previously requested the honeypot path)
http-request silent-drop if { sc_get_gpt(0,0) gt 0 }
# Protect all paths except /robots.txt, /.well-known/*, and /favicon.ico
acl unprotected_path path -m reg ^/(robots.txt|\.well-known/.*|favicon\.ico|_challenge)$
# Matches the default config of anubis of triggering on "Mozilla"
acl protected_ua hdr(User-Agent) -m beg Mozilla/
# Set stick table index 0 to 1 if request is for honeypot path
http-request sc-set-gpt(0,0) 1 if { path -m beg /blokmeplz/ }
http-request silent-drop if { path -m beg /blokmeplz/ }
acl accepted sc_get_gpt(1,0) gt 0
http-request return status 200 content-type "text/html; charset=UTF-8" hdr "Cache-control" "max-age=0, no-cache" lf-file /usr/local/etc/haproxy/challenge.html if !unprotected_path protected_ua !accepted
use_backend challenge if { path -m beg /_challenge }
http-response set-header Strict-Transport-Security "max-age=16000000; includeSubDomains; preload;"
default_backend main
backend challenge
mode http
option http-buffer-request
# The parameter to table must match the stick table used in the frontend.
http-request track-sc0 src table www
acl challenge_req method POST
http-request set-var(txn.tries) req.body_param(tries)
http-request set-var(txn.ts) req.body_param(ts)
http-request set-var(txn.host) hdr(Host),host_only
http-request set-var(txn.hash) src,concat(;,txn.host,),concat(;,txn.ts,),concat(;,txn.tries),sha2,hex
acl ts_recent date,neg,add(txn.ts) ge -60
# 4 is the difficulty, should match "diff" in challenge.html.
acl hash_good var(txn.hash) -m reg 0{4}.*
http-request sc-set-gpt(1,0) 1 if challenge_req ts_recent hash_good
http-request return status 200 if challenge_req hash_good
http-request return status 400 content-type "text/html; charset=UTF-8" hdr "Cache-control" "max-age=0" string "Bad request" if !challenge_req OR !hash_good
backend main
mode http
balance leastconn
option httpchk GET /health
http-check expect status 200
server badblocks-personal-site badblocks-personal-site:4321 check resolvers docker resolve-prefer ipv4 init-addr none
resolvers docker
nameserver dns1 127.0.0.11:53
resolve_retries 3
timeout resolve 1s
timeout retry 1s
hold valid 10s
hold obsolete 30s

View file

@ -69,6 +69,9 @@ DisallowAITraining: /
Content-Usage: ai=n Content-Usage: ai=n
Allow: / Allow: /
User-Agent: *
Disallow: /blokmeplz/
Sitemap: ${sitemapURL.href} Sitemap: ${sitemapURL.href}
`; `;