METHODS AND SYSTEMS FOR AUTOMATED SUSTAINABILITY AND MANAGEMENT OF A CLOUD INFRASTRUCTURE
Automated computer-implemented methods and systems for automated detection and termination of idle objects executing in a cloud infrastructure. The methods and systems learn rules from previous instances in which the object was terminated based on log messages associated with the previous instances. The rules are used to perform real time detection of idle instances of the object and, in response, terminate the object.
1 . A computer-implemented process that detects and terminates idle objects executing in a cloud infrastructure, the process comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous terminated instances of the object;
training rules that detect idle instances of the object based on frequencies of event types of the log messages that correspond to the previous terminated instances of the object;
displaying a graphical user interface (“GUI”) that enables a user to verify and change conditions of the rules and store the conditions in a rules database; and
automatically terminating the object in response to frequencies of event types of log messages recorded in a current time interval and satisfy conditions of a rule stored in the rules database.
2 . The process of claim 1 wherein training the rules to detect idle instances of the object comprises:
for each time interval in time intervals that correspond to previous terminated instances of the object,
determining the event type of each log message,
determining the frequency of occurrence of each event type in the time interval, and
forming a frequency distribution of the event types in the time interval from the frequency of occurrence, the frequency distribution corresponding to a terminated instance of the object; and
using a rule learning engine to generate the rules that detect idle instances of the object based on the frequency distributions that corresponding to terminated instances of the object.
3 . The process of claim 1 further comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous alive instances of the object; and
training rules that detect alive instances of the object based on frequencies of event types of the log messages that correspond to the previous alive instances of the object.
4 . The process of claim 1 wherein displaying the GUI comprises:
for each rule,
displaying one or more conditions of the rule in the GUI;
terminating one or more conditions of the rule selected for deletion by a user via the GUI; and
adding one or more conditions created by the user to the rule via the GUI.
5 . A computer system for detecting and terminating idle objects executing in a cloud infrastructure, the computer system comprising:
a display screen;
one or more processors;
one or more data-storage devices; and
machine-readable instructions stored in the one or more data-storage devices that when executed using the one or more processors control the system to perform operations comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous terminated instances of the object;
training rules that detect idle instances of the object based on frequencies of event types of the log messages that correspond to the previous terminated instances of the object;
displaying a graphical user interface (“GUI”) that enables a user to verify and change conditions of the rules and store the conditions in a rules database; and
automatically terminating the object in response to frequencies of event types of log messages recorded in a current time interval and satisfy conditions of a rule stored in the rules database.
6 . The computer system of claim 6 wherein training the rules to detect idle instances of the object comprises:
for each time interval in time intervals that correspond to previous terminated instances of the object,
determining the event type of each log message,
determining the frequency of occurrence of each event type in the time interval, and
forming a frequency distribution of the event types in the time interval from the frequency of occurrence, the frequency distribution corresponding to a terminated instance of the object; and
using a rule learning engine to generate the rules that detect idle instances of the object based on the frequency distributions that corresponding to terminated instances of the object.
7 . The computer system of claim 6 further comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous alive instances of the object; and
training rules that detect alive instances of the object based on frequencies of event types of the log messages that correspond to the previous alive instances of the object.
8 . The computer system of claim 6 wherein displaying the GUI comprises:
for each rule,
displaying one or more conditions of the rule in the GUI;
terminating one or more conditions of the rule selected for deletion by a user via the GUI; and
adding one or more conditions created by the user to the rule via the GUI.
9 . A non-transitory computer-readable medium having instructions encoded thereon for enabling one or more processors of a computer system to perform operations comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous terminated instances of the object;
training rules that detect idle instances of the object based on frequencies of event types of the log messages that correspond to the previous terminated instances of the object;
displaying a graphical user interface (“GUI”) that enables a user to verify and change conditions of the rules and store the conditions in a rules database; and
automatically terminating the object in response to frequencies of event types of log messages recorded in a current time interval and satisfy conditions of a rule stored in the rules database.
10 . The medium of claim 9 wherein training the rules to detect idle instances of the object comprises:
for each time interval in time intervals that correspond to previous terminated instances of the object,
determining the event type of each log message,
determining the frequency of occurrence of each event type in the time interval, and
forming a frequency distribution of the event types in the time interval from the frequency of occurrence, the frequency distribution corresponding to a terminated instance of the object; and
using a rule learning engine to generate the rules that detect idle instances of the object based on the frequency distributions that corresponding to terminated instances of the object.
11 . The medium of claim 9 further comprising:
retrieving log messages stored in a log file of the object with time stamps in time intervals that correspond to previous alive instances of the object; and
training rules that detect alive instances of the object based on frequencies of event types of the log messages that correspond to the previous alive instances of the object.
12 . The medium of claim 9 wherein displaying the GUI comprises:
for each rule,
displaying one or more conditions of the rule in the GUI;
terminating one or more conditions of the rule selected for deletion by a user via the GUI; and
adding one or more conditions created by the user to the rule via the GUI.