Machine learning to predict container failure for data transactions in distributed computing environment

Patent metadata

Field	Value
Publication number	US-11550672-B1
Application number	US-202117470366-A
Country	US
Kind code	B1
Filing date	Sep 9, 2021
Priority date	Sep 9, 2021
Publication date	Jan 10, 2023
Grant date	Jan 10, 2023

Abstract

Inflight transactions having predictable pod failure in distributed computing environments are managed by integrating a transaction manager into pods having containers running applications in a distributed computing environment, wherein the transaction manager records a transaction log having data indicative of historical pod failure. A pod health check that is also integrated into the pods determines predictive pod failure scenarios from the data of historical pod failure in the transaction log. Pod health can be tracked using the pod health checker by matching the predictive pod failure scenarios to transaction calls. Calls may be sent to a load balancer for recovery of pod failure for transaction calling match the predictive pod failure scenarios. Pods can be configured recover for the predictive pod failure.

First claim

What is claimed is:

1. A computer-implemented method is provided for managing inflight transactions with predicted pod failures in distributed computing environments comprising:
integrating a transaction manager into pods having containers running applications in a distributed computing environment, wherein the transaction manager records a transaction log having data indicative of historical pod failure;
integrating a pod health checker into the pods having containers running applications in a distributed computing environment, wherein the pod health checker determines predictive pod failure scenarios from the data of historical pod failure in the transaction log;
tracking pod health using the pod health checker matching the predictive pod failure scenarios to transaction calls;
sending calls for recovery of pod failure for the transaction calls match the predictive pod failure scenarios predicted to a load balancer in the distributed computing environment; and
configuring pods to recover for the predictive pod failure with the load balancer.

2. The computer-implemented method of claim 1 , wherein the transaction log is stored in block chain memory.

3. The computer-implemented method of claim 1 , wherein the pod health checker tracks pod health at intervals during an entirety of the running of the distributed computing environment.

4. The computer implemented method of claim 1 , wherein the pod health checker employs a machine learning engine for determining scenarios for the predicted pod failure from the transaction logs.

5. The computer implemented method of claim 1 , wherein the pod health checker employs a machine learning engine matching the scenarios for predicted pod failure to transaction calls for applications.

6. The computer implemented method of claim 1 , wherein the pod health checker employs a machine learning engine matching the scenarios for extracting pod recovery data from the transaction log for the matching scenarios of predictive pod failure to transaction calls.

7. The computer implemented method of claim 6 , wherein the load balancer includes a POD recovery monitor module that will route the pod recovery data to a recovery manager container in a recovery pod, wherein the recovery pod recovers the transaction that resulted in predictive pod failure.

8. The computer-implemented method of claim 1 , wherein the distributed computing environment is a cloud computing environment.

9. A system for managing inflight transactions with predicted pod failures in distributed computing environments comprising:
a hardware processor; and
a memory that stores a computer program product, which, when executed by the hardware processor, causes the hardware processor to:
integrate a transaction manager into pods having containers running applications in a distributed computing environment, wherein the transaction manager records a transaction log having data indicative of historical pod failure;
integrate a pod health checker into the pods having containers running applications in a distributed computing environment, wherein the pod health checker determines predictive pod failure scenarios from the data of historical pod failure in the transaction log;
track pod health using the pod health checker matching the predictive pod failure scenarios to transaction calls;
send calls for recovery of pod failure for the transaction calls match the predictive pod failure scenarios predicted to a load balancer in the distributed computing environment; and
configure pods to recover for the predictive pod failure with the load balancer.

10. The system of claim 9 , wherein the transaction log is stored in block chain memory.

11. The system of claim 9 , wherein the pod health checker tracks pod health at intervals during an entirety of the running of the distributed computing environment.

12. The system of claim 9 , wherein the pod health checker employs a machine learning engine for determining scenarios for the predicted pod failure from the transaction logs.

13. The system of claim 9 , wherein the pod health checker employs a machine learning engine matching the scenarios for predicted pod failure to transaction calls for applications.

14. The system of claim 9 , wherein the pod health checker employs a machine learning engine matching the scenarios for extracting pod recovery data from the transaction log for the matching scenarios of predictive pod failure to transaction calls.

15. The system of claim 14 , wherein the load balancer includes a POD recovery monitor module that will route the pod recovery data to a recovery manager container in a recovery pod, wherein the recovery pod recovers the transaction that resulted in predictive pod failure.

16. The system of claim 9 , wherein the distributed computing environment is a cloud computing environment.

17. A computer program product for managing inflight transactions with predicted pod failures in distributed computing environments comprising a computer readable storage medium having computer readable program code embodied therewith, the program instructions executable by a processor to cause the processor to:
integrate, using a processor, a transaction manager into pods having containers running applications in a distributed computing environment, wherein the transaction manager records a transaction log having data indicative of historical pod failure;
integrate, using a processor, a pod health checker into the pods having containers running applications in a distributed computing environment, wherein the pod health checker determines predictive pod failure scenarios from the data of historical pod failure in the transaction log;
track, using a processor, pod health using the pod health checker matching the predictive pod failure scenarios to transaction calls;
send, using a processor, calls for recovery of pod failure for the transaction calls match the predictive pod failure scenarios predicted to a load balancer in the distributed computing environment; and
configure, using a processor, pods to recover for the predictive pod failure with the load balancer.

18. The computer program product of claim 17 , wherein the pod health checker tracks pod health at intervals during an entirety of the running of the distributed computing environment.

19. The computer program product of claim 17 , wherein the pod health checker employs a machine learning engine for determining scenarios for the predicted pod failure from the transaction logs.

20. The computer program product of claim 17 , wherein the pod health checker employs a machine learning engine matching the scenarios for predicted pod failure to transaction calls for applications.

Assignees

• Kyndryl Inc

Inventors

• Kurian John
• Moothoor Jithesh
• Gokavarapu Nageswararao Venkata
• Srinivasan Raghavendran

Classifications

• G06F11/202

  where processing functionality is redundant (redundant communication control functionality G06F11/2005, redundant storage control functionality G06F11/2089) · CPC title

• G06F2009/4557

  Distribution of virtual machine instances; Migration and load balancing · CPC title

• G06F11/0709

  in a distributed system consisting of a plurality of standalone computer nodes, e.g. clusters, client-server systems · CPC title

• G06F9/45558

  Hypervisor-specific management and integration aspects · CPC title

• G06F11/203

  using migration · CPC title