How InterfaceRCA’s Multimodal Analysis Accurately Identifies Root Causes in Microservice Systems

Written by

in

InterfaceRCA combines system logs, numerical metrics, and distributed execution traces from microservice interfaces with a Graph Attention Network autoencoder to pinpoint interface-level root causes. On public benchmarks and real-world production datasets, it achieves an average HR@1 of 92.57 percent and NDCG@1 of 95.51 percent, with an absolute improvement of approximately 2.93 percentage points in HR@1 over FaaSRCA. The results were published in Cluster Computing.

Source:BIOENGINEER.ORG · Read original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *