Introduction - The objective of this article is to assess the completeness of data within the framework of the surveillance of COVID-19 clusters established in May 2020 by Santé publique France, and to review the surveillance data in order to improve it. Methods - The study focuses on clusters recorded between May 11 and October 28, 2020, in the Monic application, in the Occitanie region (5.6 million inhabitants). The completeness rate of the variables was analyzed by month of cluster recording, by type of community, and by severity level. The characteristics of the clusters were described when data completeness was sufficient. Results - The total number of clusters was 701, with a peak of 363 in October. Data completeness declined starting in August, particularly for the variables "size of the exposed population," "deaths," and "hospitalizations," as well as for variables used to assess severity, and starting in September for those used to assess status. Clusters in medical-social institutions and services (ESMS) were the most frequent, followed by clusters in school, university, and workplace settings. Conclusion - The surveillance system made it possible to characterize the clusters, albeit with certain limitations related to data collection quality. In the future, it would be beneficial to ensure reliable collection of variables essential for assessing severity, to streamline the system during periods of high viral circulation, and to characterize the circumstances of transmission.