Network abuse
Here is my log of spotted and reported network abuse incidents.
It started as private notes aiming to keep track of those being
fixed, and to block the hosts if they keep spamming. I decided
to make it public, since there is no private information in it
(though I'm omitting the bits I may discover that aren't public,
such as server administrator email addresses), and it may be of
interest for people trying to decide whether reporting is
worthwhile.
Spam messages
Below are incidents with spam messages that got through the
usual filters: dates, hosts, the abuse contact and other report
information, other notes.
XMPP
- 2021-09-12, 188.243.192.232, abuse@sknt.ru: no response and
spam kept coming, submitted a JabberSPAM blacklist PR.
- 2021-09-12, 138.201.50.174, stian@barmen.nu: replied that he
will investigate. Probing from ether@jabber.no.
- 2021-09-12, 54.36.115.48, info@xmpp.gg and abuse@ovh.net: no
reply from either and the spam kept coming, submitted a
blacklist PR. Probing from ink@jabber.gg.
- 2022-08-25, 138.201.25.9, abuse@hetzner.com followed
by Hetzner abuse reporting form. Subscription requests and
OMEMO-encrypted messages, similar ones from multiple services
and JIDs, with occasional plaintext being just silly. This one
is from klassic@isgeek.info. Those kept coming for at least a
month.
- 2022-08-25, 185.146.232.56, vesselwave@protonmail.com: they
deleted the user and started looking more closely for
spammers. From klassic@satisprivacy.org.
- 2022-08-25, 95.168.217.72, support@jabbim.zendesk.com and
abuse@superhosting.cz (since the first one had no
effect). From multiks@jabbim.sk.
- 2022-09-06, 170.187.181.190, abuse@linode.com. From
multiks@rows.im.
- 2022-09-10, 86.250.242.174. Did not notice at first, and
then it ceased. Probing (presence subscription requests) from
multiks@im.azurs.fr.
- 2022-10-01, 89.147.108.127, info@outerrealm.net on
2022-10-06, within 30 minutes received a reply saying that it
will be looked into, and apparently it was solved. From
ehf@msg.outerrealm.net: subscription requests at first, an odd
message saying "Request Subscription" (followed by
opportunistic OTR's whitespaces, similarly to some of the past
spammy/probing messages) on 2022-10-06.
- 2022-10-18, 78.72.102.36. Have not reported, but then it
disappeared; possibly somebody else did. From swe@qwik.space,
a subscription request.
- 2022-10-18, 78.72.102.36. Same as above: haven't reported,
but then it disappeared. From basik@qwik.space, a subscription
request.
- 2022-11-01, 138.201.50.174, stian@barmen.nu. From
floki@jabber.no: "Hi there, free for chat?". Then a
subscription request from the same JID arrived on
2023-01-03.
- 2022-11-16, 138.201.25.9, the Hetzner reporting form (since
have not found aministrator contact information). Received an
acknowledgement on 2023-01-11, a reply from the XMPP server
aministrator on 2023-01-13 saying that it doesn't look like
spam; described the issue in more detail, another reply saying
that it sounds like "complete nonsense" and suggesting to use
iptables. Asked on operators@muc.xmpp.org to ensure that my
approach is sensible, and replied to abuse@hetzner.com, asking
about their policy on XMPP spam; no reply, as of
2023-05-05. Unexpected presence subscription request and no
message (likely probing) from basik@isgeek.info.
- 2022-12-13, 138.201.50.174, stian@barmen.nu. Then again on
2023-03-08 (after an additional message from the same XMPP
address). From prtship@jabber.no/_, a presence subscription
request, and a "Hi, Free for chat?" message 3 months
later.
- 2023-01-18, 167.179.180.180, abuse@octothorn.com (on
2023-01-19). Received a reply on 2023-02-15, mentioning that
the user is being kicked off, and the account had more than
1000 contacts in the roster, most of which were pending a
subscription approval. From aus@jabber.octothorn.com/_, a
presence subscription request. The last one arrived on
2023-01-31.
Email
- 2021-02-09, 103.66.105.237, noc@cmjainimpex.in.
- 2021-03-31, 205.201.133.233, abuse@mailchimp.com.
- 2021-06-24, 2a00:1450:4864:20::641, Gmail abuse reporting
form. Apparently reporting didn't work, nothing happened
on "submit".
- 2021-06-25, 91.223.3.194, admin@skynode.pl.
- 2022-04-25, 146.19.173.107, abuse@ipconnect.services.
- 2022-04-28, 5.181.80.128, noc@4vendeta.com.
- 2022-05-29, 200.93.248.119, rolfex@powerfast.net.
- 2022-05-30, 193.218.204.206, abuse@heficed.com. The client
replied that it was solved a long time ago.
- 2022-05-31, 2607:f8b0:4864:20::e41, Gmail abuse reporting
form.
- 2022-06-30, 211.100.47.38. A Chinese ISP, probably not worth
reporting, Blacklisted
in
postscreen_access.cidr
.
- 2022-08-15, 159.183.196.221, abuse@sendgrid.com.
- 2022-11-01, 2607:5500:3000:1176::2,
support@hostwinds.com.
- 2023-05-05, 106.75.10.112, ipas@cnnic.cn. From
ucmail25.sendcloud.io.
- 2023-05-30, 69.12.91.126, abuse@quadranet.com.
- 2023-06-16, 117.50.66.12, ipas@cnnic.cn. From
ucmail17.sendcloud.io, added
sendcloud.io REJECT
spammers
into the file referenced by
postfix's check_client_access. dnswl.org returned
127.0.15.0 for it, reported it to them as spam.
- 2023-06-22, 192.119.65.137, abuse@hostwinds.com. Their mail
server (Gmail) rejects messages with the spam message
attached, reported without an attachment.
- 2023-07-21, 220.133.13.91,
hostmaster@twnic.net.tw. According to the received mail
headers, it originated from 185.225.74.219.
- 2023-09-15, 46.17.43.50, noc@baxet.ru. With valid SPF for
tiaohu.net: apparently a Chinese organization's domain name,
but a Russian hoster's IP address. Quickly received a reply
saying "Blocked" from support@justhost.asia.
- 2023-09-15, 2607:f8b0:4864:20::935, Gmail abuse reporting form.
- 2023-09-22, 2607:f8b0:4864:20::72c, Gmail abuse reporting
form. Same address as the previous one
(polachek@squadhelp.co), a follow-up.
- 2023-09-23, 2607:f8b0:4864:20::72a, Gmail abuse reporting
form. Same address as the previous two, the spammer claimed it
is the last message.
- 2023-09-25, 2607:f8b0:4864:20::f29, Gmail abuse reporting
form. A new subdomain, polachekg@go.squadhelp.co, but
continuation of the previous 3, and Gmail does nothing;
blacklisted the domain in postfix (check_sender_access).
- 2023-10-19, 209.85.128.177, Gmail abuse reporting form. From
masonlambert190@gmail.com
- 2023-11-01, 209.85.128.172, Gmail abuse reporting form. From
katherinesophia523@gmail.com
- 2023-12-05, 31.192.235.11, abuse@profitserver.ru. Phishing,
envelope-from abuse@q03.1cooldns.com, with valid DKIM and
SPF.
- 2023-12-11, 31.192.237.60, abuse@profitserver.ru. Phishing
again, envelope-from abuse@origin.1cooldns.com.
- 2023-12-11, 209.85.219.180, Gmail abuse reporting form. From
haileyjtanner@gmail.com, asking to add a link to some
furniture selling website (which supposedly has a blog post on
astronomy) from my "links" page.
- 2023-12-18, 209.85.128.170, Gmail abuse reporting form. From
haileyjtanner@gmail.com again, Gmail does not seem to do much
about outgoing spam.
- 2023-12-19, 31.192.239.9, abuse@profitserver.ru. Phishing
yet again, envelope-from=no-replies@batixtaneve.com this
time. Blacklisted 31.192.232.0/21.
- 2023-12-26, 209.85.128.169, Gmail abuse reporting form. From
haileyjtanner@gmail.com yet again, Gmail still does
nothing. Blacklisted the address in postfix
(check_sender_access).
- 2024-02-29, 204.152.197.177, abuse@quadranet.com. Spam about
electric bicycles
- 2024-03-12, 185.218.100.84, abuse@ipxo.com.
- 2024-03-18, 194.53.136.174, abuse@virtono.com. Spam about
electric bicycles, same as on 2024-03-12.
- 2024-03-20, 104.223.121.26, abuse@quadranet.com. Same as the
last two, and as on 2024-02-29: e-bikes.
- 2024-04-25, 2024-04-26, 216.9.224.143,
abuse@dchost.com. Scam, 3 messages. And one more message from
the misconfigured mail server, notifying about a failed
delivery (the "from" address matched the "to" address).
- 2024-05-09, 173.249.144.124, abuse@liquidweb.com. Posing as
a Docusign notification.
- 2024-06-12, 193.188.192.139, abuse@pipenet.hu.
- 2024-07-31, 47.90.198.34, abuse@alibaba-inc.com.
- 2024-08-08, 103.224.90.82, abuse@nexcess.net. Phishing
- 2024-09-23, 208.234.3.27, abuse@verizon.net,
abuse@ait.com. A scam, as described in "Beware of Chinese
Domain Scams" or "Chinese domain registration emails". Verizon
pointed to AIT.com, I wrote there, the "support ticket" was
closed quickly without a comment.
- 2024-09-24, 2a00:1450:4864:20::42b, Gmail abuse reporting
form. From saracody9@gmail.com, a request to link some
irrelevant website from mine.
- 2024-10-27, 219.134.170.101,
anti-spam@chinatelecom.cn. Router advertisements.
- 2024-11-18, 46.23.108.219, abuse@bullethost.net. Electric
bicycle advertisement.
- 2024-11-19, 192.154.230.159,
abuse@host4yourself.com. Electric bicycle advertisement.
- 2024-11-22, 181.214.99.201, abuse@ipxo.com. E-bikes.
- 2024-11-30, 188.127.247.224, abuse@smartape.net
(though SmartApe is reported to be a Russian hosting for
cybercriminals itself). Probing.
- 2024-12-01, 120.241.40.88, abuse@chinamobile.com. Spam about
shipping from China.
- 2024-12-04, 91.193.18.13,
abuse@hostzealot.com. E-bikes.
- 2024-12-06, 181.214.99.132,
report@abuseradar.com. E-bikes.
- 2024-12-10, 84.32.41.141,
report@abuseradar.com. E-bikes.
- 2024-12-13, 162.250.189.12, complaints@servarica.com. The
ticket was automatically created and automatically closed
without response in 36 hours; blacklisted its subnet
in
postscreen_access.cidr
.
General observations
A lot of network abuse (spam, vulnerability scans, brute-force
attacks) comes from China, plenty from Russia as well. As a side
note, Chinese researchers similarly spam the world with
fabricated research papers (though apparently they try to combat
it, up to a death penalty for researchers who commit fraud if it
harms people). Apparently wider agreements, policies, and
cultures help to fight network abuse about as well as
technological methods do. I think it is okay to rate-limit
regional IP address blocks (as described in the private server
setup notes), but not to block them completely: there may be
non-abusive users once in a while, and it would be unfair to
them. And then there are large mail providers, particularly
Gmail, not caring much about outgoing spam, while blocking them
is a bad option, given the number of legitimate users: the
ham-to-spam ratio is less than 1, but more than 0.