In a cloud service, detecting problems can only be done with synthetic transaction software simulating a user accessing those services in the public, because you can't place any hardware or software in the public cloud infrastructure environment to monitor performance. Application performance issues may be related to a slew of components such as servers, networking and storage bottlenecks, as well as the Internet between you and the public cloud. It's not work spending much money there.
So, we really are talking about APM in a private cloud environment. More specifically, we are dealing with private data centers that have virtualized server environments and automated service provisioning--that may also require the provisioning of application data collection, as well.
Many products that capture transactions like dynaTrace, Precise TPM and others can be deployed in a private cloud as an additional virtual machine, or co-provisioned along with the application. Since these applications are at the heart of the data center, they have the availability to detect and report root cause problems and properly alert the operations team when they have problems.
Increasing revenue, improving user satisfaction and maintaining performance drive any type of management tool and APM is no different. With budget limitations and fears of complex APM solutions, many organizations are looking for ways to show quick wins and extend the reasons that organizations move to the cloud in the first place.