Automated resiliency analysis in distributed systems

A method, comprising:

• determining, by a resiliency analyzer, a plurality of computing resources associated with a system architecture;
• determining, by the resiliency analyzer, one or more similar system architectures with respect to the system architecture, wherein respective identifiers of the one or more similar system architectures are stored in a repository;
• determining, by the resiliency analyzer, one or more behaviors associated with the system architecture based at least in part on the plurality of computing resources and the one or more similar system architectures identified in the repository;
• and
• generating, by the resiliency analyzer, one or more resiliency tests associated with the one or more behaviors, wherein the one or more resiliency tests are performed on the system architecture, and wherein one or more effects of the one or more behaviors on a resiliency of the system architecture are determined based at least in part on the one or more resiliency tests.

One or more non-transitory computer-readable storage media storing program instructions that, when executed on or across one or more processors, perform:

• determining a plurality of computing resources associated with a system architecture;
• determining, in a repository of system architectures, one or more similar system architectures with respect to the system architecture;
• determining one or more anticipated behaviors associated with the system architecture based at least in part on the plurality of computing resources and the one or more similar system architectures determined within the repository;
• and
• generating one or more resiliency tests associated with the one or more anticipated behaviors, wherein the one or more resiliency tests perform fault injection for a plurality of faults in the system architecture, and
• wherein the one or more resiliency tests determine a resiliency of the system architecture to the one or more anticipated behaviors.