SYSTEMS AND METHODS FOR CORRELATING DERIVED METRICS FOR SYSTEM ACTIVITY

摘要

A system can correlate derived metrics for system activity to determine problems and recommend solutions. Using a hierarchy of derived metrics from a set of raw metrics, a system can identify a problem, correlate related metrics and determine a recommended solution. For example, raw metrics can be collected about computing resources. Analyzers can process the raw metrics and outputs from other analyzers to gather metrics that include metrics derived from other metrics. When a problem symptom is discovered, derived metrics (and other metrics) can be correlated with the symptom to help identify the problem. Using the correlated metrics, a system can recommend a solution to an identified problem.

主权项

1. A system for monitoring a data center comprising:
a set of data center resources that provide one or more metrics about their operation; a hierarchy of analyzers configured to process an input of metric data and provide an output of metric data, wherein a first subset from a first level of the hierarchy of analyzers is configured to receive first metric data based on the one or more metrics, and wherein a second subset from another level of the hierarchy of analyzers is configured to receive second metric data output from one or more analyzers from a lower level in the hierarchy of analyzers; a monitoring system configured to:
detect a problem based at least in part on an output of an analyzer;correlate a plurality of analyzer outputs based at least in part on the problem; anddetermine a recommended action based at least in part on the correlated analyzer outputs; and a rendering system configured to transmit display information that highlights the identified problem, the correlated analyzer outputs and the recommended action.