Resiliency testing of applications and compute infrastructures
A computer-implemented method, according to one embodiment, includes: intentionally causing faults to be injected in a compute infrastructure, and determining whether the injected faults cause application failures. Weights are also assigned to the injected faults based on severity of the respective application failures. The weighted faults are compared, and changes to the compute infrastructure are recommended based on the comparison. Moreover, the changes that are recommended are configured to prevent the application failures. Other systems, methods, and computer program products are described in additional embodiments.
1 . A computer-implemented method, comprising:
intentionally causing faults to be injected in a compute infrastructure;
determining whether the injected faults cause application failures;
assigning weights to the injected faults based on severity of the respective application failures, wherein a negative or zero weight is assigned to ones of the injected faults that cause a change to compute infrastructure behavior, and do not cause an application error;
comparing the weighted faults;
using an artificial intelligence (AI) model that has been trained to generate physical and/or logical changes to the compute infrastructure based at least in part on the comparison, the injected faults, and/or performance of the compute infrastructure and application(s) in response to the faults being injected, wherein the physical and/or logical changes are configured to prevent the application failures; and
causing the physical and/or logical changes to be implemented in the compute infrastructure.
2 . The computer-implemented method of claim 1 , wherein assigning weights to the injected faults includes:
creating a heat map having entries identifying whether respective ones of the injected faults cause application errors;
determining weights for the injected faults identified in the heat map as causing application errors; and
correlating the determined weights with the respective injected faults identified in the heat map as causing the application errors.
3 . The computer-implemented method of claim 2 , wherein the weights are determined based at least in part on an amount of time the application experienced respective errors, wherein the weights increase as the amount of time the compute infrastructure experiences respective errors increases.
4 . The computer-implemented method of claim 3 , wherein the injected faults include infrastructure faults that intentionally subject the compute infrastructure to strain.
5 . The computer-implemented method of claim 2 , wherein the weights are determined based on the outcome and distribution difference of microservice internal calls.
6 . The computer-implemented method of claim 1 ,
wherein the AI model is trained to generate the physical and/or logical changes to the compute infrastructure based on the comparison, the injected faults, and the performance of the compute infrastructure and application in response to injecting the faults.
7 . The computer-implemented method of claim 1 , wherein the injected faults are selected based on faults previously injected in the compute infrastructure, wherein the injected faults relate to parameters of the compute infrastructure, the parameters being selected from the group consisting of: system errors, fault codes, test workloads, application topologies, and key performance indicators.
8 . A computer program product, comprising a computer readable storage medium having program instructions embodied therewith, the program instructions readable by a processor, executable by the processor, or readable and executable by the processor, to cause the processor to:
intentionally cause faults to be injected in a compute infrastructure;
determine whether the injected faults cause application failures;
assign weights to the injected faults based on severity of the respective application failures, wherein a negative or zero weight is assigned to ones of the injected faults that cause a change to compute infrastructure behavior, and do not cause an application error;
compare the weighted faults;
use an artificial intelligence (AI) model that has been trained to generate physical and/or logical changes to the compute infrastructure based at least in part on the comparison, the injected faults, and/or performance of the compute infrastructure and application(s) in response to the faults being injected, wherein the physical and/or logical changes are configured to prevent the application failures; and
cause the physical and/or logical changes to be implemented in the compute infrastructure.
9 . The computer program product of claim 8 , wherein assigning weights to the injected faults includes:
creating a heat map having entries identifying whether respective ones of the injected faults cause application errors;
determining weights for the injected faults identified in the heat map as causing application errors; and
correlating the determined weights with the respective injected faults identified in the heat map as causing the application errors.
10 . The computer program product of claim 9 , wherein the weights are determined based at least in part on an amount of time the application experienced respective errors.
11 . The computer program product of claim 10 , wherein the weights increase as the amount of time the compute infrastructure experiences respective errors increases.
12 . The computer program product of claim 9 , wherein the weights are determined based on the outcome and distribution difference of microservice internal calls.
13 . The computer program product of claim 8 , wherein
the AI model is trained to generate the physical and/or logical changes to the compute infrastructure based on the comparison, the injected faults, and the performance of the compute infrastructure and application in response to injecting the faults.
14 . The computer program product of claim 8 , wherein the injected faults are selected based on faults previously injected in the compute infrastructure, wherein the injected faults are selected from the group consisting of: system errors, fault codes, test workloads, application topologies, and key performance indicators.
15 . The computer program product of claim 14 , wherein the injected faults relate to parameters of the compute infrastructure, the parameters being selected from the group consisting of: system errors, fault codes, test workloads, application topologies, and key performance indicators.
16 . A system, comprising:
a processor; and
logic integrated with the processor, executable by the processor, or integrated with and executable by the processor, the logic being configured to:
intentionally cause faults to be injected in a compute infrastructure;
determine whether the injected faults cause application failures;
assign weights to the injected faults based on severity of the respective application failures, wherein a negative or zero weight is assigned to ones of the injected faults that cause a change to compute infrastructure behavior, and do not cause an application error;
compare the weighted faults;
use an artificial intelligence (AI) model that has been trained to generate physical and/or logical changes to the compute infrastructure based at least in part on the comparison, the injected faults, and/or performance of the compute infrastructure and application(s) in response to the faults being injected, wherein the physical and/or logical changes are configured to prevent the application failures; and
cause the physical and/or logical changes to be implemented in the compute infrastructure.
17 . The system of claim 16 , wherein
the AI model is trained to generate the physical and/or logical changes to the compute infrastructure based on the comparison, the injected faults, and the performance of the compute infrastructure and application in response to injecting the faults.