Kubernetes Vulnerability Allowed Malicious Control of Nodes
This critical vulnerability allowed any user, even unauthenticated ones, to take control of any compute node in a Kubernetes cluster.
Join the DZone community and get the full member experience.Join For Free
Major cloud providers warned customers this week of a severe vulnerability in Kubernetes that could allow remote attackers to take over entire compute nodes, potentially stealing data or corrupting production applications. Patches for the newly-discovered flaw, which affects the Kubernetes API server, have been released for the affected versions.
"With a specially crafted request, users that are authorized to establish a connection through the Kubernetes API server to a backend server can then send arbitrary requests over the same connection directly to that backend, authenticated with the Kubernetes API server’s TLS credentials used to establish the backend connection," Kubernetes developers warned on GitHub. The flaw allowed any user, even unauthenticated ones, to escalate their own privileges and take control of clusters.
Software architect Darren Shepherd found and reported the vulnerability, known as CVE-2018-1002105. Tech news sites are calling this vulnerability the first major security flaw in Kubernetes, the most popular container orchestration platform for the cloud.
No one has reported an exploit of this flaw yet — but it would be impossible to know, in any case. Since any malicious API calls wouldn't be recognized as such, they would be untraceable. Industry experts stressed the importance of prevention measures like monitoring tools, which can help find unauthorized changes. "Organizations publishing APIs for public consumption should carefully select design and technical controls to protect against known threats, including anti-automation, and far better monitoring to detect breaches," said Andrew van der Stock, senior principal consultant at Synopsys. "APIs are designed to be called after all, and when they function without errors, monitoring cannot just be of failed attempts, but also include threshold breaches around extensive and sustained access to sensitive records or changes to configuration."
Opinions expressed by DZone contributors are their own.