Method And System For The On-Demand Generation Of Graph-Like Models Out Of Multidimensional Observation Data

Patent metadata

Field	Value
Publication number	US-2022358023-A1
Application number	US-202217733105-A
Country	US
Kind code	A1
Filing date	Apr 29, 2022
Priority date	May 7, 2021
Publication date	Nov 10, 2022
Grant date	—

Abstract

Technologies are disclosed for the automated, rule-based generation of models from arbitrary, semi-structured observation data. Context data of received observation data, like data describing the location of on which a phenomenon was observed, is used to identify related observations, to generate entities in a model describing the observed data and to assign observations to model data. Mapping rules may be used for the on-demand generation of models, and different sets of mapping rules may be used to generate different models out of the same observation data for different purposes. Further, observation time data may be used to observer the temporal evolution of the generated model. Possible use cases of the so generated models include the interpretation of observation data that describes unexpected operation conditions in view of the generated model, or to determine how a monitored system reacts on changing conditions, like increased load.

First claim

What is claimed is:

1 . A computer-implemented method for monitoring performance in a distributed computing environment, comprising:
receiving, by a data ingestion module, an instance of performance monitoring data; extracting, by the data ingestion module, a performance metric from the instance of performance monitoring data in accordance with data ingestion rules, where the performance metric includes a value for the performance metric and a timestamp at which the performance metric was observed and a metric type for the performance metric; extracting, by the data ingestion module, context dimension data from the instance of performance monitoring data in accordance with the data ingestion rules, where the context dimension data identifies a given computing entity to which the performance metric pertains to and includes a key that identifies a specific context dimension and a value for the specific context dimension; creating, by the data ingestion module, a datapoint using the performance metric and the context dimension data, where the datapoint includes the value for the performance metric, the timestamp at which the performance metric was observed, the metric type for the performance metric and the context dimension data; storing, by the data ingestion module, the datapoint in a data store; receiving, by a model extraction module, a model generation request, where the model generation request identifies or contains model update rules; retrieving, by the model extraction module, datapoints from the data store in accordance with the model generation request; generating, by the model extraction module, a model element for a topology model based on the retrieved datapoints and the model update rules, where context dimension data in the retrieved data points matches an applicability criteria defined by the model update rules, and the topology model represents at least a portion of the distributed computing environment and defines relationships between computing entities in the distributed computing environment; updating, by the model extraction module, the topology model with the model element; and analyzing, by an analysis module, datapoints in the data store using the updated topology model.

2 . The method of claim 1 further comprises capturing the performance monitoring data by an agent instrumented into a computing entity in the distributed computing environment.

3 . The method of claim 1 further comprises extracting a second performance metric from the instance of performance monitoring data, where the second performance metric differs from the performance metric.

4 . The method of claim 1 further comprises receiving, by the data ingestion module, multiple instances of performance monitoring data;
extracting, by the data ingestion module, one or more performance metrics from each instance of the multiple instances of performance monitoring data in accordance with data ingestion rules; and
storing, by the data ingestion module, a subset of datapoints from the one or more performance metrics in the data store as a timeseries, where datapoints in the subset of datapoints have matching metric types and matching context dimension data.

5 . The method of claim 1 wherein the model generation request specifies at least one of a time range in which performance metrics were observed and values for context dimension data of interest, such that the datapoints are retrieved from the datastore using the at least one of the time range in which performance metrics were observed and the values for context dimension data of interest.

6 . The method of claim 1 wherein generating a model element for a topology model based on the retrieved datapoints and the model update rules further comprises one of creating a structural entity in the topology model or creating a relationship between two structural entities in the topology model.

7 . The method of claim 1 wherein the model update rules are comprised of a series of rules, such that each rule in the series of rules can be applied in parallel to the datapoints in the data store.

8 . The method of claim 1 wherein the model generation request specifies an observation period in which performance metrics were observed and further comprises retrieving datapoints from the data store that fall within the observation period, where the retrieved datapoints pertain to a particular computing entity;
creating a structural entity in the topology model for the particular computing entity; and
assigning an existence period to the particular computing entity based on timestamps from the retrieved datapoints, where the existence period indicates presence of the particular computing entity in the distributed computing environment.

9 . The method of claim 1 wherein the model generation request specifies an observation period in which performance metrics were observed and further comprises
retrieving datapoints from the data store that fall within the observation period, where the retrieved datapoints pertain to a particular relationship between computing entities;
creating a relationship in the topology model for the particular relationship; and
assigning an existence period to the relationship based on timestamps from the retrieved datapoints, where the existence period indicates presence of the particular relationship between computing entities in the distributed computing environment.

10 . The method of claim 1 wherein analyzing datapoints in the data store using the updated topology model further comprises
a) analyzing, by an analysis module, datapoints in the data store to identify a particular anomaly in a given performance metric;
b) determining, by an analysis module, a given computing entity in the distributed computing environment upon which the particular anomaly was observed;
c) creating, by an analysis module, a node for the given computing entity in a topology model in accordance with the model update rules, where the topology model represents at least a portion of the distributed computing environment and defines relationships between computing entities in the distributed computing environment;
d) identifying, by an analysis module, additional computing entities having a relationship with the given computing entity; and
e) for each additional computing entity, updating, by an analysis module, the topology model in accordance with the model update rules.
f) for each additional computing entity, searching in the topology model for one or anomalies that may have caused the particular anomaly;
g) for each of the one or more anomalies, analyzing a newly detected anomaly to determine a causal relationship with the particular anomaly; and
for each newly detected anomaly having a causal relationship with the particular anomaly, recursively performing steps c) through h).

11 . A system for monitoring performance in a distributed computing environment, comprising:
a data ingestion module configured to receive an instance of performance monitoring data, extract a performance metric from the instance of performance monitoring data in accordance with data ingestion rules and extract context dimension data from the instance of performance monitoring data in accordance with the data ingestion rules, where the performance metric includes a value for the performance metric and a timestamp at which the performance metric was observed and a metric type for the performance metric and the context dimension data identifies a given computing entity to which the performance metric pertains to and includes a key that identifies a specific context dimension and a value for the specific context dimension; wherein the date ingestion module

Assignees

• Dynatrace Llc

Inventors

• Moser Herwig
• Carpella Martin
• Ertl Otmar

Classifications

• G06F11/3447 Primary

  Performance evaluation by modeling · CPC title

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• G06F2201/835

  Timestamp · CPC title

• G06F11/3476

  Data logging (G06F11/14, G06F11/2205 take precedence) · CPC title

• G06F16/288

  Entity relationship models · CPC title