As more companies move to cloud native, it is increasingly attracting the attention of attackers, according to a new report from Cloud Native security firm Aqua.
In its 2022 Cloud Native Threat report, Aqua’s Team Nautilus analyzed attacks over the past year, noting that attackers “are…always on the lookout for easy targets—and Kubernetes is becoming such a target.”
While cryptomining was the most common malware observed in 2021, attackers are increasingly turning to backdoors and worms, the report said. Backdoors accounted for 54% of attacks, up 9 percentage points from 2020, and worms were observed in 51% of attacks, a 10 percentage-point increase, Aqua said.
Other more sophisticated activity involved rootkits, fileless execution, and loading kernel modules.
“In cloud native environments, rootkits executed on the host can be used to hide malicious processes and to reduce the chances of detection after attackers have escaped from the containerized environment,” Aqua said.
Attackers move from Docker to Kubernetes and the CI/CD pipeline
While misconfigured Docker APIs were still a popular target for attackers, Aqua saw a small decline in those attacks as threat actors began to target vulnerable CI/CD (continuous integration/continuous delivery) and Kubernetes environments.
“Kubernetes’ broad attack surface proved particularly enticing to threat actors,” Aqua said, noting that the number of malicious container images that targeted Kubernetes environments was up by 10 percentage points in 2021, to 19%.
Traditional attack tactics aimed at cloud native
In 2021, Aqua’s Team Nautilus observed attackers using traditional tactics such as password cracking and exploiting vulnerabilities in applications in an effort to steal cloud metadata.
“The implication of this activity is that attackers are using more techniques specifically reserved for applications that are running in cloud environments. This development suggests that large botnets such as Mirai, Muhstik, and Kinsing have evolved,” said Aqua.
“As the popularity of Kubernetes has risen, so too has the severity and frequency of attacks on them, as cybercriminals have realised that Kubernetes can be vulnerable,” said Steve Judd, Solutions Architect at Jetstack, a Venafi company.
“The cybercrime group TeamTNT has been a real exponent of this, having compromised more than 50,000 Kubernetes clusters over the last few years, spreading malware at will, and eventually launching a cryptominer,” Judd said.
"With innovation in the cloud increasing, so too is the number of machine identities in use. Many of these will be spun up and down in a matter of seconds. However, each needs to be given an identity, which must be managed throughout its lifecycle. Enterprises are struggling to issue and manage these identities at cloud speed and scale. The result is new security risks due to mismanagement of machine identities,” according to Judd.
Zero trust is crucial for protecting organizations against attacks targeting Kubernetes. And it’s important for businesses to not blindly trusting everything within their build environments and instead adopt a stance whereby every component of the build pipeline is proactively challenged, according to Judd.
“Automated machine identity is crucial to ensuring companies don’t kill the speed of development whilst deploying this zero trust model. Through automation, organisations can ensure the dynamic nature of cloud-native environments remain secure, as manually checking the provenance of every component of a build pipeline would take weeks. Developers need solutions that enable – instead of hinder – speed and security," Judd said.
Venafi’s machine identity platform meets the customers where they are in the journey as they rollout cloud native applications. Imagine an overarching platform with several modules that caters to various use cases. Whether you are an organization that has applications deployed in your data center or use a cloud operating model or deploy applications on Kubernetes (in a data center or in a CSP’s environment), Venafi has a module for that.
As adoption of Kubernetes matures within the organization, we’ve seen service meshes being introduced. A real world challenge with machine identity management is how to integrate with existing enterprise solutions and ensure that the identities for the mesh workloads are rooted in the enterprise chain of trust. Venafi has been addressing this challenge for the most security conscious organizations.