The subject technology retrieves information related to a set of instances of compute service managers, each instance of a particular compute service manager being associated with a set of virtual warehouses. The subject technology filters the information to determine a set of candidates from the set of instances of compute service managers. The subject technology sorts the set of candidates based at least in part on each workload of each of the set of candidates. The subject technology selects a candidate compute service manager from the set of instances of compute service managers to issue a query restart by randomly selecting an execution node, the execution node being included in a particular virtual warehouse associated with the candidate compute service manager, the selecting facilitating improving utilization of cluster resources and improving query execution on the selected candidate compute service manager.

Techniques for executing machine learning (ML) models including receiving an indication to run an ML model on a processing core; receiving a static memory allocation for running the ML model on the processing core; determining that a layer of the ML model uses more memory than the static memory allocated; transmitting, to a shared memory, a memory request for blocks of the shared memory; receiving an allocation of the requested blocks; running the layer of the ML model using the static memory and the range of memory addresses; and outputting results of running the layer of the ML model.

Disclosed are various embodiments for optimizing hypervisor paging. A hypervisor can save a machine page to a swap device, the machine page comprising data for a physical page of a virtual machine allocated to a virtual page for a process executing within the virtual machine. The hypervisor can then catch a page fault for a subsequent access of the machine page by the virtual machine. Next, the hypervisor can determine that the physical page is currently unallocated by the virtual machine in response to the page fault. Subsequently, the hypervisor can send a command to the swap device to discard the machine page saved to the swap device in response to a determination that the physical page is currently unallocated by the virtual machine.

Controlling data locality in a Kubernetes computing environment by establishing a Kubernetes computing environment including a controller and at least one executor pod for running an application, and receiving a request for a task to be run in the Kubernetes computing environment. The controller dispatches a sidecar to collect resource data from the at least one executor pod for an input to a directed acyclic graph (DAG) feature analyzer. The directed acyclic graph (DAG) feature analyzer identifies from the at least one executor pod a best dynamic resource that are available to execute. The at least one executor pod meeting the best dynamic resource that is available executes the task to be run in the Kubernetes computing.

Shared resource interference detection techniques are described. In an example, a resource detection module supports techniques to quantify levels of interference through use of working set sizes. The resource detection module selects working set sizes. The resource detection module then initiates execution of code that utilizes the shared resource based on the first working set size. The resource detection module detects a resource consumption amount based on the execution of the code. The resource detection module then determines whether the detected resource consumption amount corresponds to the defined resource consumption amount for the selected working set size.

Techniques discussed herein relate to managing service provider resources. The techniques may include receiving a first request to organize a first workload and a second workload into a space. The first workload may be associated with a first computing resource of a first service provider and the second workload may be associated with a second computing resource of a second service provider. The techniques may import data associated with the first workload and the second workload into the space. The techniques may cause an action to be performed for the first workload and the second workload by implementing a first workflow for the first workload and implementing a second workflow for the second workload.

Example embodiments employ a selective memory swapping system for selectively placing non-volatile memory devices of a computer system offline, e.g., for background updating, and online, for use by a computer system, whereby the background updating process includes a mechanism for performing forensics analysis and updating of offline memory devices while an alternate memory device is usable by a user of the first computer system.

Concepts and technologies are disclosed herein for providing a network virtualization policy management system. An event relating to a service can be detected. A first policy that defines allocation of hardware resources to host the virtual network functions can be obtained, as can a second policy that defines deployment of the virtual network functions to the hardware resources. The hardware resources can be allocated based upon the first policy and the virtual network functions can be deployed to the hardware resources based upon the second policy.

Techniques are described for filtering and normalizing training data used to build a predictive auto scaling model used by a service provider network to proactively scale users' computing resources. Further described are techniques for identifying collections of computing resources that exhibit suitably predictable usage patterns such that a predictive auto scaling model can be used to forecast future usage patterns with reasonable accuracy and to scale the resources based on such generated forecasts. The filtering of training data and the identification of suitably predictable collections of computing resources are based in part on autocorrelation analyses, and in particular on “delayed” autocorrelation analyses, of time series data, among other techniques described herein.

Embodiments of this application relate to the field of communications technologies, and provide a system service timeout processing method and an apparatus. The method includes: when a target system service thread in at least one system service thread times out, determining, by a terminal, a first application process communicating with the target system service thread, where the timeout of the target system service thread includes at least one of the following: a locked object occupied by the target system service thread is not released within a preset time, and the target system service thread is blocked; and ending, by the terminal, the first application process.