Abstract: One embodiment provides a method, including: producing, for each of a plurality of containers, a resource profile for each thread in each of the plurality of containers; identifying, for each of the plurality of containers and from, at least in part, the resource profiles, container dependencies between threads on a single of the plurality of containers; determining service dependencies between threads across different of the plurality of containers; scheduling, based upon the container dependencies and the service dependencies, threads to cores, wherein the scheduling is based upon minimizing thread processing times; and publishing the container dependencies and the service dependencies on a registry of the node clusters.

Abstract: An ensemble of autoencoder models can be trained using different seeds. The trained ensemble of autoencoder models can be run on new time series data to generate a prediction associated with the new time series data. The new time series data can include multiple dimensions per time step. Reconstruction errors can be determined for the prediction. Dimensions having highest reconstruction errors can be selected among the multiple dimensions based on a threshold. The prediction can be segmented based on bursts of the reconstruction errors over time, where temporal segments can be obtained. At least one common pattern including a set of dimensions among the selected dimensions across the temporal segments can be obtained to represent a failure fingerprint.

Abstract: An ensemble of autoencoder models can be trained using different seeds. The trained ensemble of autoencoder models can be run on new time series data to generate a prediction associated with the new time series data. The new time series data can include multiple dimensions per time step. Reconstruction errors can be determined for the prediction. Dimensions having highest reconstruction errors can be selected among the multiple dimensions based on a threshold. The prediction can be segmented based on bursts of the reconstruction errors over time, where temporal segments can be obtained. At least one common pattern including a set of dimensions among the selected dimensions across the temporal segments can be obtained to represent a failure fingerprint.

Abstract: One embodiment provides a method, including: producing, for each of a plurality of containers, a resource profile for each thread in each of the plurality of containers; identifying, for each of the plurality of containers and from, at least in part, the resource profiles, container dependencies between threads on a single of the plurality of containers; determining service dependencies between threads across different of the plurality of containers; scheduling, based upon the container dependencies and the service dependencies, threads to cores, wherein the scheduling is based upon minimizing thread processing times; and publishing the container dependencies and the service dependencies on a registry of the node clusters.

Abstract: A computer program product comprises a computer readable storage medium having program instructions embodied therewith. The program instructions are executable by one or more processors to cause the one or more processors to collect, at a network interface controller associated with a first node, a set of usage metrics from one or more other network interface controllers respectively associated with one or more other nodes. Further, the program instructions are executable by one or more processors to cause the one or more processors to select, at the network interface controller associated with the first node, one of the one or more other nodes to allocate a virtualized workload for execution, wherein the selection utilizes at least a portion of the set of usage metrics collected from each of the one or more other network interface controllers.