Abstract: Outage detection in a cloud based service is provided using usage data based error signals. Usage data is collected from component of the cloud based service or client devices of the cloud based service based on customer actions on the cloud based service. The usage data is aggregated and normalized to generate an error signal from errors generated from a component of the cloud based service. An outage is detected from the error signal. An alert that includes information associated with the outage and one or more customers impacted by the outage is generated.

Abstract: Usability of a cloud based service is recovered from a system failure. A customer transaction associated with the customer experience is executed to simulate the customer experience in the cloud based service. A failure associated with a subsystem the cloud based service is detected from an output of the customer transaction. A recovery action is determined to be associated with the failure. The recovery action is executed on the subsystem and monitored to determine a success status.

Abstract: Outage detection in a cloud based service is provided using usage data based error signals. Usage data is collected from component of the cloud based service or client devices of the cloud based service based on customer actions on the cloud based service. The usage data is aggregated and normalized to generate an error signal from errors generated from a component of the cloud based service. An outage is detected from the error signal. An alert that includes information associated with the outage and one or more customers impacted by the outage is generated.

Abstract: Usability of a cloud based service is recovered from a system failure. A customer transaction associated with the customer experience is executed to simulate the customer experience in the cloud based service. A failure associated with a subsystem the cloud based service is detected from an output of the customer transaction. A recovery action is determined to be associated with the failure. The recovery action is executed on the subsystem and monitored to determine a success status.

Abstract: The overall health of an online service is determined using health information that is obtained from the individual components of the online service. For example, an overall health score for the online service may be determined by computing an average score using health scores that are received from the different components. The amount an individual component's health score affects the overall health score of the online service is configurable. In some cases, a health alert may be sent to the component(s) experiencing the problem as well as sending the health alert to the system level. In other cases, the alerts to the component level may be suppressed and the system level administrator receives the alert. A health alert may show information such as the overall health score, the health score for each individual component, a description of any detected problems, a graphical health map of the system, and the like.