As multi-tier cloud applications become pervasive, we need better tools for
understanding their performance. This paper presents a system that analyzes
observed or desired changes to end-to-end latency prole in a large distributed
application, and identifies their underlying causes. It recognizes changes to
system conguration, workload, or performance of individual services that lead to
the observed or desired outcome. Experiments on an industrial datacenter
demonstrate the utility of the system.