Patent attributes
A technology is described for detecting a failure of a distributed system component. An example method may include registering a declarative file that may identify a distributed computing cluster in a service provider environment and provide failure criteria used to detect a failure of a distributed system component included in the distributed computing cluster. Distributed system components included in the distributed computing cluster may then be identified using information included in the declarative file. A distributed system component included in the distributed computing cluster may then be queried according to query criteria provided by the declarative file and a failure state of the distributed system component included in the distributed computing cluster may be identified based in part on a result of querying the distributed system component.