Procedures for the detection of changes in panel data typically concern offline procedures, i.e., all observations are available at the beginning of the statistical analysis and testing is done retrospectively. The present paper deals with online procedures for detecting a change in the mean of panel data arriving sequentially, with the total number of observations eventually received being random. The procedures are developed using principles of offline detection procedures [5] in combination with those used in online procedures [8]. Some limit properties of the proposed procedures are presented, as well as simulation results exhibiting desirable finitesample properties. The focus is on high-dimensional panel models, which are high dimensional time series with special properties often encountered in econometrics and financial applications.