Resource management for software containers using container profiles
A method comprises receiving a request to execute an instance of a given software container, determining source code entities of source code of the given software container, and generating a given software container profile for the given software container based at least in part on rankings associated with the source code entities. The method also comprises creating a resource management plan for the given software container utilizing one or more machine learning algorithms, the resource management plan comprising resource management metric thresholds determined based at least in part on historical resource utilization data for additional software containers having associated software container profiles similar to the given software container profile. The method further comprises executing the instance of the given software container on a software container platform, and managing resource utilization of the executing instance of the given software container based at least in part on the resource management plan.
1. A method comprising steps of:
receiving a request to execute an instance of a given software container;
determining a set of source code entities of source code of the given software container;
generating a given software container profile for the given software container based at least in part on rankings associated with the set of source code entities of the source code of the given software container;
identifying, utilizing one or more machine learning algorithms, one or more additional software containers having associated software container profiles exhibiting at least a threshold level of similarity to the given software container profile;
determining, based at least in part on historical resource utilization data for the identified one or more additional software containers, a set of resource management metric thresholds;
creating a resource management plan for the given software container, the resource management plan comprising the determined set of resource management metric thresholds;
executing the instance of the given software container on a software container platform; and
managing resource utilization of the executing instance of the given software container based at least in part on the resource management plan for the given software container;
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.
2. The method of claim 1 wherein the request to execute the instance of the given software container comprises a container image for the given software container.
3. The method of claim 1 wherein the request to execute the instance of the given software container comprises source code for one or more applications to be packaged into a container image for the given software container.
4. The method of claim 1 wherein the set of source code entities comprise at least one of software classes, software methods and software functions.
5. The method of claim 1 , wherein a given one of the rankings associated with a given one of the source code entities characterizes functional importance to an enterprise system.
6. The method of claim 5 wherein functional importance to the enterprise system is determined based at least in part on at least one of usage of the given source code entity and connectedness of the given source code entity.
7. The method of claim 6 wherein the usage of the given source code entity is determined based at least in part by analyzing basis path hits at runtime of the given source code entity by instances of software containers comprising the given source code entity.
8. The method of claim 6 wherein the connectedness of the given source code entity is determined based at least in part by identifying a number of edges connecting a node representing the given source code entity to one or more other nodes in a source code entities graph, the source code entities graph comprising nodes representing a plurality of source code entities and edges connecting the nodes representing linkages between the plurality of source code entities in a plurality of software containers.
9. The method of claim 1 wherein the software container profiles associated with the identified one or more additional software containers have respective software container profile weight values within a designated threshold of a given software container profile weight of the given software container profile.
10. The method of claim 9 wherein the designated threshold comprises k and identifying the one or more additional software containers having associated software container profiles exhibiting at least the threshold level of similarity to the given software container profile comprises utilizing a k nearest neighbor machine learning algorithm to identify k software container profiles from a repository of software container profiles, the k software container profiles comprising those software container profiles in the software container profile repository which comprise software container profile weights closest to the given software container profile weight.
11. The method of claim 10 wherein the determined set of resource management metric thresholds comprise at least one of:
a central processing unit (CPU) resource utilization limit calculated based on historical CPU resource utilization by instances of software containers associated with the k software container profiles;
a memory resource utilization limit calculated based on historical memory resource utilization by the instances of the software containers associated with the k software container profiles; and
a storage space utilization limit calculated based on historical storage space utilization by the instances of the software containers associated with the k software container profiles.
12. The method of claim 9 wherein a distance between a first software container profile weight for a first software container and a second software container profile weight for a second software container indicates relative functional importance between at least a first application of the first software container and at least a second application of the second software container.
13. The method of claim 1 wherein managing resource utilization of the executing instance of the given software container based at least in part on the resource management plan for the given container comprises:
monitoring resource utilization by the executing instance of the given software container; and
dynamically altering resource utilization by the executing instance of the given software container responsive to the monitored resource utilization being outside one or more of the determined set of resource management metric thresholds.
14. The method of claim 13 wherein dynamically altering resource utilization by the executing instance of the given software container comprises at least one of:
allocating additional resources of the container platform to the executing instance of the given software container;
de-allocating resources of the container platform from the executing instance of the given software container; and
throttling resource usage by the executing instance of the given software container.
15. A computer program product comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes the at least one processing device to perform steps of:
receiving a request to execute an instance of a given software container;
determining a set of source code entities of source code of the given software container;
generating a given software container profile for the given software container based at least in part on rankings associated with the set of source code entities of the source code of the given software container;
identifying, utilizing one or more machine learning algorithms, one or more additional software containers having associated software container profiles exhibiting at least a threshold level of similarity to the given software container profile;
determining, based at least in part on historical resource utilization data for the identified one or more additional software containers, a set of resource management metric thresholds;
creating a resource management plan for the given software container, the resource management plan comprising the determined set of resource management metric thresholds;
executing the instance of the given software container on a software container platform; and
managing resource utilization of the executing instance of the given software container based at least in part on the resource management plan for the given software container.
16. The computer program product of claim 15 wherein a given one of the rankings associated with a given one of the source code entities characterizes functional importance to an enterprise system, and wherein functional importance to the enterprise system is determined based at least in part on at least one of usage of the given source code entity and connectedness of the given source code entity.
17. The computer program product of claim 15 wherein:
the software container profiles associated with the identified one or more additional software containers have respective software container profile weight values within a designated threshold of a given software container profile weight of the given software container profile;
the designated threshold comprises k; and
identifying the one or more additional software containers having associated software container profiles exhibiting at least the threshold level of similarity to the given software container profile comprises utilizing a k nearest neighbor machine learning algorithm to identify k software container profiles from a repository of software container profiles, the k software container profiles comprising those software container profiles in the software container profile repository which comprise software container profile weights closest to the given software container profile weight.
18. An apparatus comprising:
at least one processing device comprising a processor coupled to a memory;
the at least one processing device being configured to perform steps of:
receiving a request to execute an instance of a given software container;
determining a set of source code entities of source code of the given software container;
generating a given software container profile for the given software container based at least in part on rankings associated with the set of source code entities of the source code of the given software container;
identifying, utilizing one or more machine learning algorithms, one or more additional software containers having associated software container profiles exhibiting at least a threshold level of similarity to the given software container profile;
determining, based at least in part on historical resource utilization data for the identified one or more additional software containers, a set of resource management metric thresholds;
creating a resource management plan for the given software container, the resource management plan comprising the determined set of resource management metric thresholds;
executing the instance of the given software container on a software container platform; and
managing resource utilization of the executing instance of the given software container based at least in part on the resource management plan for the given software container.
19. The apparatus of claim 18 wherein a given one of the rankings associated with a given one of the source code entities characterizes functional importance to an enterprise system, and wherein functional importance to the enterprise system is determined based at least in part on at least one of usage of the given source code entity and connectedness of the given source code entity.
20. The apparatus of claim 18 wherein:
the software container profiles associated with the identified one or more additional software containers have respective software container profile weight values within a designated threshold of a given software container profile weight of the given software container profile;
the designated threshold comprises k; and
identifying the one or more additional software containers having associated software container profiles exhibiting at least the threshold level of similarity to the given software container profile comprises utilizing a k nearest neighbor machine learning algorithm to identify k software container profiles from a repository of software container profiles, the k software container profiles comprising those software container profiles in the software container profile repository which comprise software container profile weights closest to the given software container profile weight.