Automated rightsizing of containerized application with optimized horizontal scaling
A method for rightsizing an application, including analyzing, via processing circuitry, metrics from containers; determining, via the processing circuitry, a resource allocation and a target resource utilization for an application workload based on the metrics; and configuring, via the processing circuitry, the application workload based on the resource allocation, the target resource utilization, and the metrics.
1 . A method for rightsizing an application, comprising:
acquiring, via processing circuitry, one or more metrics from containers in the application;
determining both a resource allocation and a target resource utilization corresponding to an application workload of the application by inputting the acquired one or more metrics to a machine-learning model, the machine-learning model being configured to simultaneously output both (1) a recommended resource allocation indicating a number of allocated resources, and (2) a recommended target resource utilization indicating a target percentage usage of each allocated resource, the recommended resource allocation and the recommended target resource utilization being determined in tandem by the machine-learning model; and
automatically configuring, via the processing circuitry, the application workload based on the determined resource allocation, the determined target resource utilization, and the acquired one or more metrics.
2 . The method of claim 1 , wherein the acquired one or more metrics are collected from deployed containers and analyzed by the machine-learning model simultaneously to determine both the resource allocation and the target resource utilization for the application workload.
3 . The method of claim 2 , further comprising receiving a configuration input from a user, wherein the machine-learning model is configured based on the configuration input.
4 . The method of claim 1 ,
wherein the determination of the target resource utilization is used for simultaneous horizontal scaling of the application workload.
5 . The method of claim 1 , further comprising generating, using machine learning, a prediction of resource usage metrics and operating costs based on the determined resource allocation and the determined target resource utilization before configuring the application workload.
6 . The method of claim 1 , further comprising continuously and automatically configuring the application workload with the determined resource allocation and the determined target resource utilization.
7 . The method of claim 1 , wherein the acquired one or more metrics input to the machine-learning model include a number of replicas, a maximum number and a minimum number of allowable replicas, and a number and a type of metrics used by a horizontal autoscaler.
8 . A device, comprising:
processing circuitry configured to:
acquire one or more metrics from containers in an application,
determine both a resource allocation and a target resource utilization corresponding to an application workload of the application by inputting the acquired one or more metrics to a machine-learning model, the machine-learning model being configured to simultaneously output both (1) a recommended resource allocation indicating a number of allocated resources, and (2) a recommended target resource utilization indicating a target percentage usage of each allocated resource, the recommended resource allocation and the recommended target resource utilization being determined in tandem by the machine-learning model, and
automatically configure the application workload based on the determined resource allocation, the determined target resource utilization, and the acquired one or more metrics.
9 . The device of claim 8 , wherein the acquired one or more metrics are collected from deployed containers and analyzed by the machine learning model simultaneously to determine both the resource allocation and the target resource utilization for the application workload.
10 . The device of claim 9 , wherein the processing circuitry is further configured to receive a configuration input from a user, wherein the machine-learning model is configured based on the configuration input.
11 . The device of claim 8 ,
wherein the determination of the target resource utilization is used for simultaneous horizontal scaling of the application workload.
12 . The device of claim 8 , wherein the processing circuitry is further configured to generate, using machine learning, a prediction of resource usage metrics and operating costs based on the determined resource allocation and the determined target resource utilization before configuring the application.
13 . The device of claim 8 , wherein the processing circuitry is further configured to continuously and automatically configure the application workload with the determined resource allocation and the determined target resource utilization.
14 . A non-transitory computer-readable storage medium for storing computer-readable instructions that, when executed by a computer, cause the computer to perform a method, the method comprising:
acquiring one or more metrics from containers in an application;
determining both a resource allocation and a target resource utilization corresponding to an application workload of the application by inputting the acquired one or more metrics to a machine-learning model, the machine-learning model being configured to simultaneously output both (1) a recommended resource allocation indicating a number of allocated resources, and (2) a recommended target resource utilization indicating a target percentage usage of each allocated resource, the recommended resource allocation and the recommended target resource utilization being determined in tandem by the machine-learning model; and
automatically configuring the application workload based on the determined resource allocation, the determined target resource utilization, and the acquired one or more metrics.
15 . The non-transitory computer-readable storage medium of claim 14 , wherein the acquired one or more metrics are collected from deployed containers and analyzed simultaneously by the machine-learning model to determine the resource allocation and the target resource utilization for the application workload.
16 . The non-transitory computer-readable storage medium of claim 15 , further comprising receiving a configuration input from a user, wherein the machine-learning model is configured based on the configuration input.
17 . The non-transitory computer-readable storage medium of claim 14 ,
wherein the determination of the target resource utilization is used for simultaneous horizontal scaling of the application workload.
18 . The non-transitory computer-readable storage medium of claim 14 , further comprising generating, using machine learning, a prediction of resource usage metrics and operating costs based on the determined resource allocation and the determined target resource utilization before configuring the application workload.
19 . The non-transitory computer-readable storage medium of claim 14 , further comprising continuously and automatically configuring the application workload with the determined resource allocation and the determined target resource utilization.