Certain aspects of the present disclosure provide techniques for handling crash events in a software application using application-agnostic machine learning models. An example method generally includes receiving a data set of crash reports from a software application for analysis. Using a first neural network, a representation of each respective crash report in the data set is generated. The data set of crash reports and a mapping between functions in the software application and a multidimensional space are input into the first neural network. Each respective crash report in the data set is classified using a second neural network and the representation of each crash report in the data set. One or more actions are taken with respect to the software application based on the classifying each respective crash report in the data set.

Approaches in accordance with various embodiments provide for the management of system event data in a computing device. In particular, various embodiments provide an intelligent persistent buffer for system event log (SEL) messages. A SEL message can be generated by system BIOS on a computing device, which can send this message over an appropriate interface to a target recipient, such as the BMC. Instead of being received directly to the BMC, however, the SEL message can be received to a logic device, such as a CPLD, that is able to analyze the message, determine that the message relates to an important system event, and can cause this message to be stored to a persistent buffer. The BMC can then subsequently request the buffered SEL message from the logic device to take an appropriate action.

Methods and apparatus relating to lane based normalized historical error counter view for faulty lane isolation and disambiguation of transient versus persistent errors are described. In an embodiment, a plurality of storage entries store error information to be detected at one or more physical lanes of an interface. Faulty lane detection logic circuitry determines which of the one or more physical lanes is faulty or more likely to be faulty based at least in part on the stored error information for the one or more physical lanes of the interface. The stored error information comprises historical error details for the one or more physical lanes of the interface. Other embodiments are also disclosed and claimed.

A Peripheral Component Interface Express (PCIe) card includes a circuit board, a device mounted on the circuit board, and a PCIe processor mounted on the circuit board. The PCIe processor is communicatively coupled to the device and a host processor of a host system. The PCIe processor is configured to detect a power signal on an auxiliary (AUX) power rail of the PCIe card. A periodic detection of a state of the device is performed based on detecting the power signal on the AUX power rail. A signal indicative of the state of the device is encoded for transmission to the host processor of the host system. PCIe link training is performed via a PCIe interface with the host system. The PCIe link training is initiated based on the signal indicative of the state of the device.

A system may include a cryptographic accelerator to generate a first check value based on a payload received in a message, and provide the first check value to a first comparator and to a second comparator. The system may include the first comparator to receive the first check value from the cryptographic accelerator, determine whether the first check value matches a second check value, the second check value being a check value received in the message, and provide a first output indicating whether the first check value matches the second check value. The system may include the second comparator to receive the first check value from the cryptographic accelerator, determine whether the first check value matches the second check value, and provide a second output indicating whether the first check value matches the second check value.

A method, computer program product, and computer system for performing, by a computing device, a check on an internal secondary device on a first node during a boot software stack initialization. It may be determined that the internal secondary device is corrupt based upon, at least in part, the check. The first node may access a recovery operating system and an image repository of an internal secondary device on a second node. The internal secondary device on the first node may be rebuilt based upon, at least in part, the recovery operating system and the image repository of the internal secondary device on the second node.

A lightweight always-on monitoring, collecting, diagnosing, and correcting utility operates in an enhanced storage manager that manages a data storage managements system. The always-on utility provides a comprehensive and pro-active approach, which is intended to reduce, if not altogether eliminate, the need for after-the-fact diagnostics. The always-on utility also enforces so-called best practices and other heuristics, which include pro-actively activating certain database settings that are not enabled by default; manipulating certain aspects of the database to improve performance; and reporting aspects that are outside best-practice parameters to the trouble report system so that system administrators and/or developers may intervene before a catastrophic failure occurs. In some cases, the best-practice parameters represent heuristics designed by the present inventors to improve the performance and general health of the management database.

Provided is an error prediction apparatus including at least one processor and at least one memory configured to receive environment data of a surveying site in which a surveying instrument is installed, to input the environment data of the surveying site into an error prediction model and predict a predicted error that occurs in a surveying result obtained by the surveying instrument under an environment of the surveying site, and to create display data for displaying the predicted error when the predicted error exceeds an allowable value. The error prediction model is a learned model created by machine learning for a surveying instrument of the same model as the surveying instrument by using a set of the environment data indicating an environment of the time of surveying and error data in a surveying result as teacher data.

Systems and methods are presented for selection of compatible components for an observed system. An exemplary method comprises collecting parameters of one or more components of the system, assessing conformity of the one or more components of the system with a required state of the system, identifying one or more anomalies based on the assessment of conformity, analyzing the one or more anomalies to identify a class and parameters of the system corresponding to the one or more anomalies, determining one or more models of methods of restoration of the system, selecting one or more components that meets requirements of the one or more models of methods of restoration and implementing the one or more components in the system that are compatible with the system to eliminate the one or more anomalies.

A computer implemented method includes receiving an exception generated based on programming code, generating exception features from the received exception, the generated exception features being generated based on a set exception features derived from search logs, and executing a machine learning model on the received exception and generated exception features to provide information from the search logs identified as most helpful to resolve the received exception, wherein the machine learning model was trained on training data comprising extracted exceptions and the set of exception features derived from the search logs.