Detecting a quality-related faulty component and predicting uncorrectable errors incurred by a component using machine learning

A computerized method comprising:

• identifying, from historical component data, a first group of components from a plurality of components that have incurred an uncorrectable error and a second group of components from the plurality of components that have not incurred the uncorrectable error;
• for each component in the first group of components:
• extracting features associated with the component, the features associated with the component comprising one or more of:
• features of the component, features of a compute node in which the component is executed on, and features of services provided by the compute node;
• and
• comparing the features associated with the component with features associated with each of the other components in the first group of components:
• based on the comparing, identifying a particular component comprising a pattern of features associated therewith that increases a probability of the particular component to incur the uncorrectable error;
• identifying a particular feature from the pattern of features that, if changed, decreases the probability of the particular component to incur the uncorrectable error;
• and
• based on the identifying, providing alternatives for the particular feature such that, when one of the alternatives is implemented, the probability of the particular component to incur the uncorrectable error is decreased.

A computer storage medium storing computer-executable instructions that, upon execution by a processor, cause the processor to perform the following:

• identifying, from historical component data, a first group of components from a plurality of components that have incurred an uncorrectable error and a second group of components from the plurality of components that have not incurred the uncorrectable error;
• for each component in the first group of components:
• extracting features associated with the component, the features associated with the component comprising one or more of:
• features of the component, features of a compute node in which the component is executed on, and features of services provided by the compute node;
• and
• comparing the features associated with the component with features associated with each of the other components in the first group of components:
• based on the comparing, identifying a particular component comprising a pattern of features associated therewith that increases a probability of the particular component to incur the uncorrectable error;
• identifying a particular feature from the pattern of features that, if changed, decreases the probability of the particular component to incur the uncorrectable error;
• and
• based on the identifying, providing alternatives for the particular feature such that, when one of the alternatives is implemented, the probability of the particular component to incur the uncorrectable error is decreased.