The invention discloses a fault prediction and self-healing method and
system for intelligent operation and maintenance of an e-commerce platform, and relates to the technical field of intelligent operation and maintenance of computers. The method comprises the following steps: collecting multi-dimensional heterogeneous
telemetry data, carrying out space-
time alignment by utilizing a link tracking identifier, and constructing a unified
feature vector containing service
semantics and resource states; inputting the unified
feature vector into a fault prediction model, and generating an early warning
signal containing a fault type and a probability by analyzing a
time sequence trend; in response to the early warning
signal, matching a target self-healing strategy in a self-healing strategy
library by using a
causal inference logic; and finally, realizing closed-loop repair through an automatic arrangement interface execution strategy. According to the method, the data island problem is solved through multi-
modal data alignment, the dynamic baseline or graph neural network is utilized to effectively distinguish the e-commerce large promotion flow and the
system fault, the safety of the self-healing decision is ensured through a
causal inference mechanism, the
false alarm rate is remarkably reduced, and the fault
recovery time is shortened.