Managing computer workloads across distributed computing clusters
Computer workloads can be managed across distributed computing clusters according to some aspects of the present disclosure. In one example, a system can receive a request from a workload manager for identifying a computing cluster to which to assign a workload. The system can determine that the workload is to be assigned to a particular computing cluster among a plurality of computing clusters based on historical information about replica deployment by the particular computing cluster. The system can then transmit a response to the workload manager for causing the workload manager to assign the workload to the particular computing cluster.
1. A system comprising:
a processor; and
a memory including instructions that are executable by the processor for causing the processor to:
receive historical information associated with a computing cluster, the historical information including a statistical distribution associated with a software component executed by the computing cluster during a prior timespan, wherein the statistical distribution indicates how many times the computing cluster executed distinct numbers of replicas of the software component during the prior timespan;
determine an assignment of a workload to the computing cluster based on a metric derived from the statistical distribution; and
cause the workload to be assigned to the computing cluster.
2. The system of claim 1 , wherein the historical information includes statistical metrics associated with replicas of the software component deployed by the computing cluster during the prior timespan.
3. The system of claim 1 , wherein the software component is a serverless function or a microservice.
4. The system of claim 1 , wherein the computing cluster is an edge cluster at a physical edge of a distributed computing environment.
5. The system of claim 1 , wherein causing the workload to be assigned to the computing cluster comprises:
transmitting an indication of the assignment to a workload manager, wherein the workload manager is configured to cause the workload to be handled by the computing cluster.
6. The system of claim 1 , wherein the statistical distribution includes a first data value indicating a first number of times in which a first number of replicas of the software component concurrently executed in the computing cluster during the prior timespan, and the statistical distribution includes a second data value indicating a second number of times in which a second number of replicas of the software component concurrently executed in the computing cluster during the prior timespan.
7. The system of claim 1 , wherein the instructions are further executable by the processor for causing the processor to:
determine an estimated resource usage by the computing cluster based on the workload and the metric; and
select the computing cluster from among a plurality of computing clusters in response to determining that the estimated resource usage is less than a resource-usage threshold.
8. The system of claim 1 , wherein the historical information includes a plurality of statistical distributions associated with a plurality of software components executed by the computing cluster during the prior timespan.
9. A method comprising:
receiving, by a processor, historical information associated with a computing cluster, the historical information including a statistical distribution associated with a software component executed by the computing cluster during a prior timespan, wherein the statistical distribution indicates how many times the computing cluster executed distinct numbers of replicas of the software component during the prior timespan;
determining, by the processor, an assignment of a workload to the computing cluster based on a metric derived from the statistical distribution; and
causing, by the processor, the workload to be assigned to the computing cluster.
10. The method of claim 9 , wherein the software component is a serverless function or a microservice.
11. The method of claim 9 , wherein the computing cluster is an edge cluster at a physical edge of a distributed computing environment.
12. The method of claim 9 , wherein causing the workload to be assigned to the computing cluster comprises:
transmitting an indication of the assignment to a workload manager, wherein the workload manager causes the workload to be handled by the computing cluster.
13. The method of claim 9 , wherein the statistical distribution includes a first data value indicating a first number of times in which a first number of replicas of the software component concurrently executed in the computing cluster during the prior timespan, and the statistical distribution includes a second data value indicating a second number of times in which a second number of replicas of the software component concurrently executed in the computing cluster during the prior timespan.
14. The method of claim 9 , further comprising:
determining an estimated resource usage by the computing cluster based on the workload and the metric; and
selecting the computing cluster from among a plurality of computing clusters in response to determining that the estimated resource usage is less than a resource-usage threshold.
15. The method of claim 9 , wherein the historical information includes a plurality of statistical distributions associated with a plurality of software components executed by the computing cluster during the prior timespan.
16. A non-transitory computer-readable medium comprising program code that is executable by a processor for causing the processor to:
receive historical information associated with a computing cluster, the historical information including a statistical distribution associated with a software component executed by the computing cluster during a prior timespan, wherein the statistical distribution indicates how many times the computing cluster executed distinct numbers of replicas of the software component during the prior timespan;
determine an assignment of a workload to the computing cluster based on a metric derived from the statistical distribution; and
cause the workload to be assigned to the computing cluster.
17. The non-transitory computer-readable medium of claim 16 , wherein causing the workload to be assigned to the computing cluster comprises:
transmitting an indication of the assignment to a workload manager, wherein the workload manager is configured to cause the workload to be handled by the computing cluster.
18. The non-transitory computer-readable medium of claim 16 , wherein the statistical distribution includes a first data value indicating a first number of times in which a first number of replicas of the software component concurrently executed in the computing cluster during the prior timespan, and the statistical distribution includes a second data value indicating a second number of times in which a second number of replicas of the software component concurrently executed in the computing cluster during the prior timespan.
19. The non-transitory computer-readable medium of claim 16 , further comprising program code that is executable by the processor for causing the processor to:
determine an estimated resource usage by the computing cluster based on the workload and the metric; and
select the computing cluster from among a plurality of computing clusters in response to determining that the estimated resource usage is less than a resource-usage threshold.
20. The non-transitory computer-readable medium of claim 16 , wherein the historical information includes a plurality of statistical distributions associated with a plurality of software components executed by the computing cluster during the prior timespan.