I manage a self-hosted Gitlab instance on a server in DigitalOcean. A few months ago, I got a notification from DigitalOcean Security Center that my instance has been disconnected from the network because it was detected performing a kind of Distributed Denial of Service (DDoS) attack.
I tried to solve the issue, I suspected that it was just a Gitlab runner process that made some requests that were falsy detected as a DDOS attack. DigitalOcean has very strict rules for this case that it becomes difficult for me to re-activate the network connection of my server before I perform a kind of recovery process and system scanning. That's why I choose to deploy a new server and back up the repository.
After I managed to deploy a new instance and re-run the repository server, I upgraded the server with the latest patch that is suitable. I also thought that the issue was only caused by falsy security detection. After that moment, I left my server untouched for months later.
Someday, I checked the administrator panel and went to the users' list. I was surprised that there are three unknown accounts with administrator level. The account names are started with "dexbcx". I deleted all the accounts spontaneously. I try to figure out what just happened. Then, I remembered the event months ago and I assume that my instance had been impaired by malicious activity.
After I browsed the internet, the event is related to the CVE-2021-22205 issue. So, the attacker can gain access to my instance by exploiting a vulnerability on Gitlab then utilize my instance to perform a DDOS attack by sending a large number of bytes to their target. Some discussion was also escalated in Gitlab Forum about this issue and how to resolve it.
Comments
Post a Comment