Proactive asset failure remediation utilizing conformal asset state prediction
An apparatus comprises a processing device configured to identify a given one of one or more assets in an information technology infrastructure associated with a support indicator associated with a priority level, and to obtain information characterizing state transitions of the one or more assets. The processing device is also configured to determine, based at least in part on a current state of the given asset identified utilizing the obtained information, a probability of the given asset transitioning to each of a plurality of states and to select, based at least in part on the determined probabilities, one of the plurality of states as a predicted future state of the given asset utilizing conformal prediction. The processing device is further configured to modify the priority level of the support indicator associated with the given asset based at least in part on the predicted future state of the given asset.
1. 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:
identifying a given one of one or more assets in an information technology infrastructure associated with a support indicator, the support indicator being associated with a priority level;
obtaining information characterizing state transitions of the one or more assets in the information technology infrastructure;
determining, based at least in part on a current state of the given asset identified utilizing the obtained information, two or more asset transition probabilities for the given asset, a first one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a first one of a plurality of states, a second one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a second one of the plurality of states different than the first one of the plurality of states, the first state being associated with a first one of two or more types of failure, and the second state being associated with a second one of the two or more types of failure, the second type of failure being different than the first type of failure, the two or more types of failure being associated with respective different failure severities;
selecting, based at least in part on the determined two or more asset transition probabilities, a given one of the plurality of states as a predicted future state of the given asset utilizing conformal prediction, the conformal prediction characterizing relative rankings of (i) the determined first asset transition probability for the given asset among a first set of possible asset transition probabilities for transitioning to the first state and (ii) the determined second asset transition probability for the given asset among a second set of possible asset transition probabilities for transitioning to the second state; and
modifying the priority level of the support indicator associated with the given asset based at least in part on a given one of the two or more types of failure associated with the predicted future state of the given asset.
2. The apparatus of claim 1 wherein the obtained information characterizing state transitions of the one or more assets in the information technology infrastructure comprises telemetry data collected from the one or more assets in the information technology infrastructure.
3. The apparatus of claim 1 wherein the obtained information characterizing state transitions of the one or more assets in the information technology infrastructure comprises application logs from one or more tools that monitor the one or more assets in the information technology infrastructure.
4. The apparatus of claim 1 wherein determining the two or more asset transition probabilities comprises determining the two or more asset transition probabilities utilizing a Markov chain.
5. The apparatus of claim 4 wherein determining the two or more asset transition probabilities comprises utilizing a transition matrix configured in accordance with the Markov chain, wherein rows of the transition matrix represent possible values of the current state of the given asset and columns of the transition matrix represent possible future states of the given asset.
6. The apparatus of claim 1 wherein determining the two or more asset transition probabilities comprises:
calculating transition counts from the current state of the given asset to each of the plurality of states in the obtained information characterizing state transitions of the one or more assets in the information technology infrastructure; and
calculating a given asset transition probability of the given asset transitioning from the current state to the given state as a ratio of the calculated transition count for the given state to a sum of the transition counts for the plurality of states.
7. The apparatus of claim 6 wherein determining the two or more asset transition probabilities comprises constructing a graph comprising nodes representing respective ones of the plurality of states and edges connecting the nodes representing respective ones of the two or more asset transition probabilities.
8. The apparatus of claim 1 wherein selecting the given state as the predicted future state of the given asset utilizes Mondrian conformal prediction.
9. The apparatus of claim 8 wherein selecting the given state as the predicted future state of the given asset utilizing Mondrian conformal prediction comprises generating a calibration table for each of two or more of the plurality of states.
10. The apparatus of claim 9 wherein the two or more of the plurality of states are associated with a same determined asset transition probability.
11. The apparatus of claim 9 wherein the calibration table for the given state comprises a ranked ordering of predicted asset transition probabilities for transitioning to the given state.
12. The apparatus of claim 9 wherein selecting the given state as the predicted future state of the given asset utilizing Mondrian conformal prediction further comprises selecting the given state having a highest confidence level among the two or more states, the highest confidence level exceeding a designated confidence threshold.
13. The apparatus of claim 12 , wherein the confidence level for the given state is determined by:
identifying a rank of the determined asset transition probability for the given state in the calibration table for the given state;
determining a number of rankings below the identified rank in the calibration table for the given state; and
calculating a ratio of the determined number of rankings below the identified rank in the calibration table for the given state to a total number of rankings in the calibration table for the given state.
14. The apparatus of claim 1 wherein modifying the priority level of the support indicator associated with the given asset comprises increasing the priority level of the support indicator responsive to the predicted future state of the given asset being a failure state.
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:
identifying a given one of one or more assets in an information technology infrastructure associated with a support indicator, the support indicator being associated with a priority level;
obtaining information characterizing state transitions of the one or more assets in the information technology infrastructure;
determining, based at least in part on a current state of the given asset identified utilizing the obtained information, two or more asset transition probabilities for the given asset, a first one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a first one of a plurality of states, a second one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a second one of the plurality of states different than the first one of the plurality of states, the first state being associated with a first one of two or more types of failure, and the second state being associated with a second one of the two or more types of failure, the second type of failure being different than the first type of failure, the two or more types of failure being associated with respective different failure severities;
selecting, based at least in part on the determined two or more asset transition probabilities, a given one of the plurality of states as a predicted future state of the given asset utilizing conformal prediction, the conformal prediction characterizing relative rankings of (i) the determined first asset transition probability for the given asset among a first set of possible asset transition probabilities for transitioning to the first state and (ii) the determined second asset transition probability for the given asset among a second set of possible asset transition probabilities for transitioning to the second state; and
modifying the priority level of the support indicator associated with the given asset based at least in part on a given one of the two or more types of failure associated with the predicted future state of the given asset.
16. The computer program product of claim 15 wherein selecting the given state as the predicted future state of the given asset utilizes Mondrian conformal prediction by:
generating a calibration table for each of two or more of the plurality of states, the calibration table for each of the plurality of states comprising a ranked ordering of predicted asset transition probabilities for transitioning to that state; and
selecting the given state having a highest confidence level among the two or more states, the highest confidence level exceeding a designated confidence threshold, the confidence level for the given state being determined by: identifying a rank of the determined asset transition probability for the given state in the calibration table for the given state; determining a number of rankings below the identified rank in the calibration table for the given state; and calculating a ratio of the determined number of rankings below the identified rank in the calibration table for the given state to a total number of rankings in the calibration table for the given state.
17. The computer program product of claim 15 wherein modifying the priority level of the support indicator associated with the given asset comprises increasing the priority level of the support indicator responsive to the predicted future state of the given asset being a failure state.
18. A method comprising:
identifying a given one of one or more assets in an information technology infrastructure associated with a support indicator, the support indicator being associated with a priority level;
obtaining information characterizing state transitions of the one or more assets in the information technology infrastructure;
determining, based at least in part on a current state of the given asset identified utilizing the obtained information, two or more asset transition probabilities for the given asset, a first one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a first one of a plurality of states, a second one of the two or more asset transition probabilities characterizing probability of the given asset transitioning to a second one of the plurality of states different than the first one of the plurality of states, the first state being associated with a first one of two or more types of failure, and the second state being associated with a second one of the two or more types of failure, the second type of failure being different than the first type of failure, the two or more types of failure being associated with respective different failure severities;
selecting, based at least in part on the determined two or more asset transition probabilities, a given one of the plurality of states as a predicted future state of the given asset utilizing conformal prediction, the conformal prediction characterizing relative rankings of (i) the determined first asset transition probability for the given asset among a first set of possible asset transition probabilities for transitioning to the first state and (ii) the determined second asset transition probability for the given asset among a second set of possible asset transition probabilities for transitioning to the second state; and
modifying the priority level of the support indicator associated with the given asset based at least in part on a given one of the two or more types of failure associated with the predicted future state of the given asset;
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.
19. The method of claim 18 wherein selecting the given state as the predicted future state of the given asset utilizes Mondrian conformal prediction by:
generating a calibration table for each of two or more of the plurality of states, the calibration table for each of the plurality of states comprising a ranked ordering of predicted asset transition probabilities for transitioning that state; and
selecting the given state having a highest confidence level among the two or more states, the highest confidence level exceeding a designated confidence threshold, the confidence level for the given state being determined by: identifying a rank of the determined asset transition probability for the given state in the calibration table for the given state; determining a number of rankings below the identified rank in the calibration table for the given state; and calculating a ratio of the determined number of rankings below the identified rank in the calibration table for the given state to a total number of rankings in the calibration table for the given state.
20. The method of claim 18 wherein modifying the priority level of the support indicator associated with the given asset comprises increasing the priority level of the support indicator responsive to the predicted future state of the given asset being a failure state.