Troubleshooting in Software-as-a-Service (SaaS) environment is inherently complex due to involvement of multiple sub-systems, which operate under lots of uncertainty. A research has been conducted to develop a Multi Agent System (MAS) for troubleshooting in SaaS environments by implementing an agent for each subsystem. Each agent has its personal ontology with the sub-system specific knowledge, whereas the domain ontology comprises of commonly accessible knowledge. In presence of an event requiring troubleshooting, the agents start communicating with each other and arrive at a globally acceptable solution.