Detecting anomalous states of machines
First Claim
1. A computer-implemented method for implementation by one or more data processors forming part of at least one computing device to facilitate detection and avoidance of undesirable system states of a system, the method comprising:
- monitoring, by the one or more data processors, states of the system, the system comprising a plurality of components, the states of the system having state types and determined based on data instances of the plurality of components and/or the system, the data instances comprising machine-log-lines;
receiving, by the one or more data processors, data instance groups representative of the states of the system, the data instance groups comprising one or more data instances associated with a state of the system and having time information;
identifying, by the one or more data processors, a representative data instance for each data instance group;
determining, by the one or more data processors, a sequence of state transitions based on the representative data instances and the time information, wherein each element of the sequence comprises a data instance group identifier and a data instance identifier;
determining, based on the sequence of state transitions, by the one or more data processors, a distribution of state types within the data instance groups to identify infrequent state types;
translating, by the one or more data processors, the sequence of state transitions into feature vectors, wherein the feature vectors comprise a first feature vector indicating a starting state of the sequence and a second feature vector indicating a time for the transition between a first state and a second state of the sequence, wherein each feature vector is associated with one or more feature classes;
calculating, by the one or more data processors, a feature score for each feature vector based on kernel density estimation;
determining, by the one or more data processors, a feature class score for individual feature classes of the one or more feature classes, the feature class score comprising a sum of feature scores having the same feature class;
calculating, by the one or more data processors, a sequence anomaly score based on the feature class scores across the sequence of state transitions, the sequence anomaly score indicating a likelihood of a rare state and/or rare sequence;
identifying, based on sequence anomaly scores, by the one or more data processors, rare states and/or rare sequences, wherein rare sequences include sequences of data instance groups that occur prior to an occurrence of a state having an infrequent state type; and
providing, by the one or more data processors and in response to identifying the rare states and/or rare sequences, a notification of the occurrence of a rare sequence on the system.
1 Assignment
0 Petitions
Accused Products
Abstract
The state of a system is determined in which data sets are generated that include a plurality of data instances representing states of one or more components of a computer system. The data instances generated by one or more data set sources that are configured to output a data instance in response to a trigger associated with the one or more components. The data instances are normalized by the application of one or more rules. The data instances from individual data set sources are separately collated to generate groups of time-specific collated data instances. State types may be assigned to each of the collated data instance groups. Distributions of state-types across the groups may be determined and a list of infrequent state-types may be generated based on the determined distributions of state-types across the groups.
-
Citations
19 Claims
-
1. A computer-implemented method for implementation by one or more data processors forming part of at least one computing device to facilitate detection and avoidance of undesirable system states of a system, the method comprising:
-
monitoring, by the one or more data processors, states of the system, the system comprising a plurality of components, the states of the system having state types and determined based on data instances of the plurality of components and/or the system, the data instances comprising machine-log-lines; receiving, by the one or more data processors, data instance groups representative of the states of the system, the data instance groups comprising one or more data instances associated with a state of the system and having time information; identifying, by the one or more data processors, a representative data instance for each data instance group; determining, by the one or more data processors, a sequence of state transitions based on the representative data instances and the time information, wherein each element of the sequence comprises a data instance group identifier and a data instance identifier; determining, based on the sequence of state transitions, by the one or more data processors, a distribution of state types within the data instance groups to identify infrequent state types; translating, by the one or more data processors, the sequence of state transitions into feature vectors, wherein the feature vectors comprise a first feature vector indicating a starting state of the sequence and a second feature vector indicating a time for the transition between a first state and a second state of the sequence, wherein each feature vector is associated with one or more feature classes; calculating, by the one or more data processors, a feature score for each feature vector based on kernel density estimation; determining, by the one or more data processors, a feature class score for individual feature classes of the one or more feature classes, the feature class score comprising a sum of feature scores having the same feature class; calculating, by the one or more data processors, a sequence anomaly score based on the feature class scores across the sequence of state transitions, the sequence anomaly score indicating a likelihood of a rare state and/or rare sequence; identifying, based on sequence anomaly scores, by the one or more data processors, rare states and/or rare sequences, wherein rare sequences include sequences of data instance groups that occur prior to an occurrence of a state having an infrequent state type; and providing, by the one or more data processors and in response to identifying the rare states and/or rare sequences, a notification of the occurrence of a rare sequence on the system. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
-
19. A system capable of detecting anomalous states of the system, comprising:
-
at least one programmable processor; and
,a non-transitory storage medium readable by at the least one processor and storing instructions which, when executed by the at least one programmable processor, implement operations comprising; monitoring states of a machine comprising a plurality of components, the states of the machine having state types and determined based on data instances of the plurality of components and/or the system, the data instances comprising machine-log-lines; receiving data instance groups representative of the states of the system, the data instance groups comprising one or more data instances associated with a state of the system and having time information; identifying a representative data instance for a data instance group; determining a sequence of state transitions based on the representative data instances and the time information, wherein an element of the sequence comprises a data instance group identifier and a data instance identifier; determining, based on the sequence of state transitions, a distribution of state types within the data instance groups to identify infrequent state types; translating the sequence of state transitions into feature vectors, wherein the feature vectors comprise a first feature vector indicating a starting state of the sequence and a second feature vector indicating a time for the transition between a first state and a second state of the sequence, wherein a feature vector is associated with one or more feature classes; calculating a feature score for the feature vector based on kernel density estimation; determining a feature class score for individual feature classes of the one or more feature classes, the feature class score comprising a sum of feature scores having the same feature class; calculating a sequence anomaly score based on the feature class scores across the sequence of state transitions, the sequence anomaly score indicating a likelihood of a rare state and/or rare sequence; identifying, based on sequence anomaly scores, rare states and/or rare sequences, wherein rare sequences include sequences of data instance groups that occur prior to an occurrence of a state having an infrequent state type; and providing, in response to identifying the rare states and/or rare sequences, a notification of the occurrence of a rare sequence on the machine.
-
Specification