An apparatus includes a memory and a processor. The processor collects from a database server, statistics indicating a performance of the server. The processor applies to the statistics, a machine learning algorithm configured to identify, based in part on the statistics, one or more issues affecting the performance of the database server. Applying the machine learning algorithm includes applying to the statistics, a regression algorithm trained to identify one or more candidate issues. Applying the machine learning algorithm also includes applying to the one or more candidate issues, a classification algorithm trained to classify each candidate issue as an issue of the one or more issues or a false positive. The processor additionally generates a script that, when executed, remediates the candidate issues of the one or more candidate issues classified by the classification algorithm as the issues of the one or more issues. The processor further executes the script.