Multi-modal artificial intelligence root cause analysis
A data processing system implements obtaining build logs that include information associated with a software build problem; analyzing the logs to generate a knowledge graph identifying the relationship between various entities in the logs; extracting a signature of a candidate root cause of the build problem from the knowledge graph representing a subset of nodes and edges of the knowledge graph; providing the signature of the candidate root cause to a graphical language model to obtain a prediction of a category of root cause failure selected from among a plurality of root cause failures; constructing a prompt for a language model to generate a root cause failure analysis that describes the root cause of the build problem, the prompt including the category of root cause; receiving the root cause failure analysis from the language model; and performing one or more actions in response to receiving the root cause failure analysis.
1 . A data processing system comprising:
a processor; and
a machine-readable storage medium storing executable instructions that, when executed, cause the processor alone or in combination with other processors to perform operations of:
obtaining build logs that include information associated with a software build that experienced a build problem;
generating a knowledge graph by analyzing the build logs using a knowledge graph generation unit, the knowledge graph identifying a relationship between various entities in the build logs;
extracting a signature of a candidate root cause of the build problem from the knowledge graph, the signature representing a subset of nodes and edges of the knowledge graph;
providing the signature of the candidate root cause to a graphical language model to obtain a prediction of a category of root cause failure, the graphical language model being trained to receive the signature of the candidate root cause and to predict a category of root cause failure from among a plurality of root cause failures based on the signature of the candidate root cause;
constructing a prompt for a language model using a prompt construction unit, the prompt instructing the language model to generate a root cause failure analysis that describes the root cause of the build problem, the prompt including the category of root cause predicted by the graphical language model;
providing the prompt as an input to the language model;
receiving the root cause failure analysis from the language model; and
performing one or more actions in response to receiving the root cause failure analysis.
2 . The data processing system of claim 1 , wherein extracting the signature of the candidate root cause of the build problem from the knowledge graph further comprises:
filtering out nodes and edges of the knowledge graph that do not appear more than a threshold number of times in the build logs.
3 . The data processing system of claim 1 , wherein the language model is a Large Language Model (LLM).
4 . The data processing system of claim 1 , wherein constructing the prompt for the language model further comprises including instructions to the language model to include a description of the root cause failure and instructions for addressing the root cause failure.
5 . The data processing system of claim 1 , wherein the machine-readable storage medium further includes instructions configured to cause the processor alone or in combination with other processors to perform operations of:
providing the root cause analysis to a moderation service to analyze the root cause analysis for potentially objectionable content; and
discarding the root cause analysis in response to the moderation service determining that the root cause analysis includes potentially objectionable content.
6 . The data processing system of claim 1 , wherein performing the one or more actions in response to receiving the root cause failure analysis further comprises:
causing a user interface of an application of a client device to present the root cause failure analysis.
7 . The data processing system of claim 1 , wherein performing the one or more actions in response to receiving the root cause failure analysis further comprises:
storing the root cause analysis and prompt in a build-related prompt datastore; and
providing a user interface for searching the build-related prompt datastore.
8 . A method implemented in a data processing system for performing a root cause analysis, the method comprising:
obtaining build logs that include information associated with a software build that experienced a build problem;
generating a knowledge graph by analyzing the build logs using a knowledge graph generation unit, the knowledge graph identifying a relationship between various entities in the build logs;
extracting a signature of a candidate root cause of the build problem from the knowledge graph, the signature representing a subset of nodes and edges of the knowledge graph;
providing the signature of the candidate root cause to a graphical language model to obtain a prediction of a category of root cause failure, the graphical language model being trained to receive the signature of the candidate root cause and to predict a category of root cause failure from among a plurality of root cause failures based on the signature of the candidate root cause;
constructing a prompt for a language model using a prompt construction unit, the prompt instructing the language model to generate a root cause failure analysis that describes the root cause of the build problem, the prompt including the category of root cause predicted by the graphical language model;
providing the prompt as an input to the language model;
receiving the root cause failure analysis from the language model; and
performing one or more actions in response to receiving the root cause failure analysis.
9 . The method of claim 8 , wherein extracting the signature of the candidate root cause of the build problem from the knowledge graph further comprises:
filtering out nodes and edges of the knowledge graph that do not appear more than a threshold number of times in the build logs.
10 . The method of claim 8 , wherein the language model is a Large Language Model (LLM).
11 . The method of claim 8 , wherein constructing the prompt for the language model further comprises including instructions to the language model to include a description of the root cause failure and instructions for addressing the root cause failure.
12 . The method of claim 8 , further comprising:
providing the root cause analysis to a moderation service to analyze the root cause analysis for potentially objectionable content; and
discarding the root cause analysis in response to the moderation service determining that the root cause analysis includes potentially objectionable content.
13 . The method of claim 8 , wherein performing the one or more actions in response to receiving the root cause failure analysis further comprises:
causing a user interface of an application of a client device to present the root cause failure analysis.
14 . The method of claim 8 , wherein performing the one or more actions in response to receiving the root cause failure analysis further comprises:
storing the root cause analysis and the prompt in a build-related prompt datastore; and
providing a user interface for searching the build-related prompt datastore.
15 . A data processing system comprising:
a processor; and
a machine-readable storage medium storing executable instructions that, when executed, cause the processor alone or in combination with other processors to perform operations of:
obtaining logs that include information associated with a problem with software;
generating a knowledge graph by analyzing the logs using a knowledge graph generation unit, the knowledge graph identifying a relationship between various entities in the logs;
extracting a signature of a candidate root cause of the problem from the knowledge graph, the signature representing a subset of nodes and edges of the knowledge graph;
providing the signature of the candidate root cause to a graphical language model to obtain a prediction of a category of root cause failure, the graphical language model being trained to receive the signature of the candidate root cause and to predict a category of root cause failure from among a plurality of root cause failures based on the signature of the candidate root cause;
constructing a prompt for a language model using a prompt construction unit, the prompt instructing the language model to generate a root cause failure analysis that describes the root cause of the problem with the software, the prompt including the category of root cause predicted by the graphical language model;
providing the prompt as an input to the language model;
receiving the root cause failure analysis from the language model; and
performing one or more actions in response to receiving the root cause failure analysis.
16 . The data processing system of claim 15 , wherein extracting the signature of the candidate root cause of the problem from the knowledge graph further comprises:
filtering out nodes and edges of the knowledge graph that do not appear more than a threshold number of times in the logs.
17 . The data processing system of claim 15 , wherein the language model is a Large Language Model (LLM).
18 . The data processing system of claim 15 , wherein constructing the prompt for the language model further comprises including instructions to the language model to include a description of the root cause failure and instructions for addressing the root cause failure.
19 . The data processing system of claim 15 , wherein the machine-readable storage medium further includes instructions configured to cause the processor alone or in combination with other processors to perform operations of:
providing the root cause analysis to a moderation service to analyze the root cause analysis for potentially objectionable content; and
discarding the root cause analysis in response to the moderation service determining that the root cause analysis includes potentially objectionable content.
20 . The data processing system of claim 15 , wherein the logs include runtime error information associated with the software.