AUTOMATED WORKLOAD MONITORING BY STATISTICAL ANALYSIS OF LOGS

Methods, computer program products, and systems are presented. The methods include, for instance: A plurality of log messages generated while an application processes two or more workloads are collected. The application is microservice-based and each of the microservices generates respective log messages. A time domain analysis is performed to lay out respective iteration counts of the microservices of the application. Keyword patterns in the log messages are compiled and stored in a keywords database. A frequency domain analysis on the keyword patterns for respective numbers of appearances in the log messages and stored in a frequency domain analysis database. Logs generated from a new workload are automatically monitored for abnormality by comparing the logs by the new workload to the statistical pattern of keywords as established by previous workloads. For a deviation greater than a threshold, an alert is generated with issue location and delivered to preselected recipients.