Use of ephemeral workloads to monitor compute environments

1 . A method comprising:

receiving, with an ephemeral job controller, a request to initiate an ephemeral job within a computing environment having one or more containers and corresponding container support functionality;

analyzing, with the ephemeral job controller, the request to initiate the ephemeral job to determine resources and/or limits based on a target container to which an ephemeral job is to be deployed;

initiating, with the ephemeral job controller, the ephemeral job corresponding to the request in the target container with limits determined from the target container;

obtaining, with the ephemeral job controller, one or more metrics from the target container via the ephemeral job;

providing, with the ephemeral job controller, the one or more metrics to a remote entity outside of the target container; and

terminating, with the ephemeral job controller and within a pre-specified period of time, the ephemeral job.

2 . The method of claim 1 , further comprising initiating and monitoring, with the ephemeral job controller, a plurality of ephemeral jobs across a corresponding plurality of containers.

3 . The method of claim 1 , wherein the ephemeral job controller is configured to support one or more of: rate limits on pods and/or Internet Protocol (IP) addresses, per-namespace resource limiting, feedback looks between orchestrator and job controller, running Spark jobs as either ephemeral or persisted jobs, multi-tenancy, intelligent scheduling, and load balancing across multiple managed service clusters.

4 . The method of claim 1 , wherein the one or more metrics are utilized to generate a polygraph to establish a baseline of behavior allowing for the future detection of deviations from that baseline.

5 . The method of claim 4 , wherein polygraph data is maintained for a set of applications in a datacenter, and such polygraph data is combined to make a datacenter view across the set of applications.

6 . The method of claim 1 , wherein the ephemeral job comprises a Spark-compliant job.

7 . The method of claim 1 , wherein the container comprises a Kubernetes-compliant container.

8 . A non-transitory computer-readable medium having stored therein instructions that, when executed by one or more hardware processors, are configurable to cause the one or more hardware processors to:

receive, with an ephemeral job controller, a request to initiate an ephemeral job within a computing environment having one or more containers and corresponding container support functionality;

analyze, with the ephemeral job controller, the request to initiate the ephemeral job to determine resources and/or limits based on a target container to which an ephemeral job is to be deployed;

initiate, with the ephemeral job controller, the ephemeral job corresponding to the request in the target container with limits determined from the target container;

obtain, with the ephemeral job controller, one or more metrics from the target container via the ephemeral job;

provide, with the ephemeral job controller, the one or more metrics to a remote entity outside of the target container; and

terminate, with the ephemeral job controller and within a pre-specified period of time, the ephemeral job.

9 . The non-transitory computer-readable medium of claim 8 further comprising instructions that, when executed by the one or more hardware processors, are configurable to cause the one or more hardware processors to:

initiate, with the ephemeral job controller, a plurality of ephemeral jobs across a corresponding plurality of containers; and

monitor, with the ephemeral job controller, the plurality of ephemeral jobs across the corresponding plurality of containers.

10 . The non-transitory computer-readable medium of claim 8 , wherein the ephemeral job controller is configured to support one or more of: rate limits on pods and/or Internet Protocol (IP) addresses, per-namespace resource limiting, feedback looks between orchestrator and job controller, running Spark jobs as either ephemeral or persisted jobs, multi-tenancy, intelligent scheduling, and load balancing across multiple managed service clusters.

11 . The non-transitory computer-readable medium of claim 8 , wherein the one or more metrics are utilized to generate a polygraph to establish a baseline of behavior allowing for the future detection of deviations from that baseline.

12 . The non-transitory computer-readable medium of claim 11 , wherein polygraph data is maintained for a set of applications in a datacenter, and such polygraph data is combined to make a datacenter view across the set of applications.

13 . The non-transitory computer-readable medium of claim 8 , wherein the ephemeral job comprises a Spark-compliant job.

14 . The non-transitory computer-readable medium of claim 8 , wherein the container comprises a Kubernetes-compliant container.

15 . A system comprising:

a memory subsystem having a plurality of memory devices; and

a set of hardware processors coupled with the memory subsystem, the set of hardare processors configurable to:

receive a request to initiate an ephemeral job within a computing environment having one or more containers and corresponding container support functionality;

analyze the request to initiate the ephemeral job to determine resources and/or limits based on a target container to which an ephemeral job is to be deployed;

initiate the ephemeral job corresponding to the request in the target container with limits determined from the target container;

obtain one or more metrics from the target container via the ephemeral job;

provide the one or more metrics to a remote entity outside of the target container; and

terminate within a pre-specified period of time, the ephemeral job.

16 . The system of claim 15 , wherein the set of hardware processors are further configurable to:

initiate a plurality of ephemeral jobs across a corresponding plurality of containers; and

monitor the plurality of ephemeral jobs across the corresponding plurality of containers.

17 . The system of claim 15 , wherein the ephemeral job controller is configured to support one or more of: rate limits on pods and/or Internet Protocol (IP) addresses, per-namespace resource limiting, feedback looks between orchestrator and job controller, running Spark jobs as either ephemeral or persisted jobs, multi-tenancy, intelligent scheduling, and load balancing across multiple managed service clusters.

18 . The system of claim 15 , wherein the one or more metrics are utilized to generate a polygraph to establish a baseline of behavior allowing for the future detection of deviations from that baseline.

19 . The system of claim 18 , wherein polygraph data is maintained for a set of applications in a datacenter, and such polygraph data is combined to make a datacenter view across the set of applications.

20 . The system of claim 15 , wherein the ephemeral job comprises a Spark-compliant job.