I have some servers I maintain and I'm the only one who should legitimately be logging in or even attempting to over Secure Shell (SSH). It doesn't stop other users from trying their hand though.
In one 24 hour period on 27 January 2019, my server had login requests using 128 different invalid user names. Some are likely just user names commonly found on Internet exposed systems. Others are looking for specific types of devices with exposed default user names that may also have default passwords.
I use key based authentication and lock down SSH, as you should, so this activity doesn't pose a threat to my infrastructure. It's still interesting to see what people are trying. I don't believe they're wasting their time with random attempts. I believe devices with default or common user names and passwords must be exposed to the Internet.
The full list of names that had been tried is at the bottom of the post in case you want to have a look and investigate any for yourself.
To extract the invalid names from the auth.log file which keeps track of every SSH login attempt among other activity related to authentication, I looked through the log for the information I was after.
Lines for attempts to access the server over SSH using an invalid user name contained the word Invalid.
I ran the following command to find lines from 27 January that have the word Invalid in them. It then grabs just the field with the user name by cutting on spaces and returning only field 8. It then does a sort on unique values so I don't have to deal with duplicates. It output the names to the file
sudo grep "Jan 27" /var/log/auth.log | grep Invalid | cut -d " " -f 8 | sort -u > usernames.txt
Here are some of the user names attempted that jumped out at me or came up in Google searches.
|Possible Associated App
|RedHat/CentOS Crash Debugger
|Default for many network devices such as Netgear
|Apache web server
|AtScale Intelligence Platform
|Git revision control system
|Self-hosted Git service
|Pivotal Greenplum database administrator
|Common Hadoop database username
|Atlassian's Jira issue and project tracking
|Possible Apache Kafka username
|Possibly related to Libsys India
|Majordomo mailing list manager
|Hadoop mapred map reduction
|Minecraft game server
|Myth TV server
|Nagios network monitoring server
|Postfix mail server
|Vagrant by HashiCorp rapid deployment manager
|RealVNC or TightVNC remote login software
|vyatta open source router software
|WebLogic Online Transaction Processing
|WordPress Content Management System (CMS)
If you don't have to expose your service to the Internet, then don't. You can protect it with a firewall and VPN to your network if you need access.
If you do have to expose it, take some precautions. To secure SSH in particular, create and use a non-root user account, permit only key based authentication, disable root remote access, enable your firewall, consider an app like Fail2ban, and consider changing from the default TCP port of 22 to a high port like 2223. Keep your systems up to date and check your logs frequently.
There are sites people use to feed automated tools that scour the Internet looking for vulnerable systems. Don't let yours fall prey. You'll notice many of the names below on that list and others like it.
I took a slightly deeper look at one login,
ubnt, on my personal blog if you're interested.
Here's the full list. That's a lot of attempts for one 24 hour period.