CausalRCA: Causal inference based precise fine-grained root cause localization for microservice applications

Effectively localizing root causes of performance anomalies is crucial to enabling the rapid recovery and loss mitigation microservice applications in cloud. Depending on granularity that can be localized, a service operator may take different actions, e.g., restarting or migrating services if only faulty localized (namely, coarse-grained) scaling resources specific indicative metrics fine-grained). Prior research mainly focuses coarse-grained localization, there now growing interest fine-grained cause localization identify metrics. Causal inference (CI) based methods have gained popularity recently for but currently used CI limitations, such as linear causal relations assumption strict data distribution requirements. To tackle these challenges, we propose framework named CausalRCA implement fine-grained, automated, real-time localization. The uses gradient-based structure learning method generate weighted graphs localize We conduct coarse- evaluate CausalRCA. Experimental results show has significantly outperformed baseline accuracy, average AC@3 metric 0.719, increase 10% compared with methods. In addition, Avg@5 improved by 9.43%. Codes are open-sourced found our Github repository