US 12,468,593 B2
Root cause identification in hybrid applications via probing
Saurabh Jha, White Plains, NY (US); Larisa Shwartz, Greenwich, CT (US); Robert Filepp, Westport, CT (US); Frank Bagehorn, Dottikon (CH); and Jesus Maria Rios Aliaga, Philadelphia, PA (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Mar. 30, 2023, as Appl. No. 18/128,370.
Prior Publication US 2024/0330096 A1, Oct. 3, 2024
Int. Cl. G06F 11/00 (2006.01); G06F 11/07 (2006.01); G06F 11/30 (2006.01)
CPC G06F 11/079 (2013.01) [G06F 11/3089 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
automatically detecting, during runtime operation of a hybrid application, an anomaly in performance of the hybrid application based on a specification of required performance and collected passive monitoring data;
generating a causal generative model based on relationships between hybrid application components and computing system architecture components extracted from the collected passive monitoring data;
executing a root cause identification (RCI) logic of an RCI engine on the causal generative model to identify a set of candidate root causes of the detected anomaly, wherein the RCI engine comprises automated reinforcement learning computer logic;
automatically and dynamically identifying one or more probes for active monitoring data collection targeting the set of candidate root causes;
executing the one or more probes on the set of candidate root causes to collect probe data;
performing reinforcement learning of the RCI engine at least by updating the RCI logic based on the probe data;
updating the set of candidate root causes based on the reinforcement learning of the RCI engine, wherein the one or more probes are executed to test at least one corresponding hybrid application component or computing system architecture component and collect operational data from the at least one corresponding hybrid application component or computing system architecture component; and
using an identified root cause as a basis for resolving the detected anomaly and to strengthen the hybrid application against one or more future anomalies of a same type as the detected anomaly.