Mastodon

Noise Watch: Week #1

Week covers data from 2026-11-07 to 2026-11-13. Tracking app is here

Here is the summary of the week.

RequestsIPs200403abort
37557790810592 / 28%950 / 3%4132 / 11%

I have implemented some new abort rules in my Caddyfile so abort has seen a spike which is expected. Very low 200 number is probably an evidence that traffic to my pages is really mostly noise rather than legitimate humans visiting. Three key statuses that I am tracking are just a tiny portion of what server replies with (most of that is 404). Requests Per Status

Another side of this is that requests are often not directed to a particular page but just probing the server for different things. Requests per Page

Top URIs requests shows this well:

URIRequestsIPs
/wp-admin/install.php?step=1321118
/.git/config23092
/.env19886
/assets/css/theme-forgejo-auto.css?v=17.0.1~gitea-1.22.0195187
/assets/js/index.js?v=17.0.1~gitea-1.22.0186176
/assets/js/webcomponents.js?v=17.0.1~gitea-1.22.0180169
/assets/css/index.css?v=17.0.1~gitea-1.22.0167158
/assets/css/repo-branch-tag-selector.d9ca1da2.css123120
/assets/js/repo-branch-tag-selector.f396185d.js115112
/.env.local11460
/api/session/properties10610
/wp-login.php8934
/.env.production8759
/api/.env8455
/.env.backup8255
/info.php8056
/app/.env7950
/.env.development7851
/.env.save7850
/.env.staging7750

Most of it is just probing for Wordpress, env files.

AI Crawlers identified by IP ranges were not very active this week, showing measly numbers accrosse the board.

CrawlerIPsAvg Time SpanRequestsPer IP
claude_crawl1513226.000000 secs15510
meta_crawl60203.605275 secs861
openai_train16817.812500 secs825
amazon_train330.000000 secs371
apple_crawl220.318199 secs362
openai_search80.000000 secs111

IPs Spike

From November 9 to November 11 there was a spike in number of IPs that visited. The main actors are displayed here:

ASNHandleEventsIPs
AS45102ALIBABA-CN-NET231102550
AS132203TENCENT-NET-AP-CN15900819
AS9808CHINAMOBILE-CN3180270
AS13335CLOUDFLARENET2625181
AS4837CHINA169-BACKBONE1860165

Multiple IP ranges were hitting me from AS45102. Most of it went to git instance where they were surfing deep through all the public repos. I cannot prove that this is all one actor. Seems like they were methodically going through all the endpoints, then found git instance.

CIDRHostRequestsIPs
47.80.0.0/13git174701947
47.78.0.0/15git5525599
47.89.128.0/17blog602
198.11.128.0/18blog302
47.80.0.0/13notes/wiki153
47.254.0.0/16blog51
47.77.128.0/17baseball app51

I could not identifiy that these were bots, not from User-Agent at least but the almost even spread between majority of those UAs probably means that it was an automated thing.

User-AgentRequests
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/153.0.0.0 Safari/537.361100
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/151.0.0.0 Safari/537.361135
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/152.0.0.0 Safari/537.361215
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/152.0.0.0 Safari/537.361175
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/148.0.0.0 Safari/537.361085
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/148.0.0.0 Safari/537.361150
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/145.0.0.0 Safari/537.361080
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0.0.0 Safari/537.361130
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/151.0.0.0 Safari/537.361265
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/150.0.0.0 Safari/537.361065
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/153.0.0.0 Safari/537.361095
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.361055
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.361205
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/144.0.0.0 Safari/537.361250
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/114.0.0.0 Safari/537.365
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/145.0.0.0 Safari/537.361250
Mozilla/5.0 (iPhone; CPU iPhone OS 13_2_3 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.0.3 Mobile/15E148 Safari/604.195
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/147.0.0.0 Safari/537.361210
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/150.0.0.0 Safari/537.361085
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/144.0.0.0 Safari/537.361065
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0.0.0 Safari/537.361205
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/147.0.0.0 Safari/537.361190

The interesting part is those abort responses the server issued. Turns out, few of those IPs were already in abort blocks in my Caddyfile (those came from Bob Rudis who once share his rules on Mastodon) but not entire IP range was covered by rules so some of them got 200 and some of them were aborted.

All public repos were properly scraped by each IP range.

CIDRRepoRequestsIPs200403Abort
47.80.0.0/13glog:public147101793470009970
47.78.0.0/15glog:public4615548165002940
47.80.0.0/13retrosheetpy:public11452141400995
47.80.0.0/13nordtheme:public8851721550730
47.80.0.0/13asnaic:public6401271350505
47.78.0.0/15retrosheetpy:public40076600340
47.78.0.0/15nordtheme:public30559550250
47.78.0.0/15asnaic:public18036600120
47.80.0.0/13other (including / and assets)105895010
47.89.128.0/17other (including / and assets)6012000
198.11.128.0/18other (including / and assets)3011000
47.78.0.0/15other (including / and assets)2512500
47.77.128.0/17other (including / and assets)51000
47.254.0.0/16other (including / and assets)51000

I have added those IPs to another abort block. There were not seemingly doing anything harmful but the scraping part is annoying.

Scanners

From the number of requests point of view these were the Top-10:

IPRequestsASN
34.187.246.401244AS396982
45.138.12.172566AS218785
20.217.186.29553AS8075
45.115.26.203491AS199457
8.234.49.30486AS3356
8.234.49.30486AS396982
207.175.218.17377AS396982
13.140.161.237342AS51167
16.192.150.116335AS16509
34.220.116.146335AS16509
43.166.206.32335AS132203
102.220.162.137320AS197769
102.133.224.63306AS8075
34.62.31.226294AS396982
34.24.153.243278AS396982
34.51.179.90278AS396982
35.228.99.210278AS396982
35.229.211.165278AS396982
35.235.93.105278AS396982
34.27.201.146277AS396982
34.51.194.104277AS396982

Checked them all on CrowdSec and the all are highlighted as malicious. In my case these were mostly doing general scanning, looking for WordPress, php files or env and other secrets. All these went into another block in Caddyfile.


The week was nothing special overall. The Noise is still out there and it still overwhelming the Signal. Ping me on Mastodon (link down in the footer) if anything here is missing or wrong. I am not an expert in traffic analysis and I am happy to learn from people smarter than me.