This work introduces a methodology for cloud accountability that assures system dependability in terms of availability and reliability. This assurance is provided relative to the cloud service level agreement. The presented methodology is guided by the NIST SP800-86 digital forensic model, that motivates the collection, examination and analysis of data from the cloud platform, and the generated evidence including logs and context are reported to appropriate cloud agents. As part of this work, we present a novel approach to collecting digital evidence to support cloud-based system dependability, using the Virtual Machine Introspection (VMI) technique. Our VMI approach complements, as well as checks the dependability metrics provided by the cloud service providers (CSPs) as evidence. This methodology, including the VMI approach is particularly relevant since it provides a means of addressing the perceived lack of trust for cloud-based services towards cloud accountability. Our research focuses on applying an evidence-based methodologycloud accountability method-to cloud-based system engineering for assuring cloud agents of the dependability of cloud platforms.