Detecting attackers who target containerized clusters

Abstract

A method for operation of a deception management server, for detecting and hindering attackers who target containerized clusters of a network, including learning the network environment, including finding existing container instances, finding existing services and relationships, extracting naming conventions in the environment, and classifying the most important assets in the environment, creating deceptions based on the learning phase, the deceptions including one or more of (i) secrets, (ii) environment variables pointing to deceptive databases, web servers or active directories, (iii) mounts, (iv) additional container instances comprising one or more of file server, database, web applications and SSH, (v) URLs to external services, and (vi) namespaces to fictional environments, planting the created deceptions via a container orchestrator, via an SSH directly to the containers, or via the container registry, and issuing an alert when an attacker attempts to connect to a deceptive entity.