Predictive Intelligence and applications that submit (ML) solutions training requests [ml_solution] can wait for several hours before they are processed by the ML Scheduler during peak periods and there is a backlog of ML training requestsSummaryWhen you submit a Predictive Intelligence solution for training, in most cases the request will be picked up and processed within a few minutes. However, during busy periods when there is a peak in these AI Service requests, it can take several hours for the ML (Machine Learning) Scheduler backlog to be processed before your Predictive Intelligence solution training request is picked up by the ML Scheduler and processed. This also applies to any application that uses the Predictive Intelligence Framework, such as Automation Discovery, Intent Discovery, Fingerprint Discovery and other applications that adds Predictive Intelligence. Unfortunately, we do not guarantee Predictive Intelligence solution training times and during busy periods, it can take longer than normal for the training to start on your Predictive Intelligence solutions. In some cases during our peak periods for these AI Service requests, it can take several hours for the backlog to clear before the solution is processed for training. There is a retry mechanism every 4 hours and eventually the training request will time out if not picked up within 12 hours. However training time outs are now less frequent and we are constantly monitoring the demand. We apologise for any inconvenience this delay in processing the training request may cause you. We are always looking to improve the performance with the ever-increasing demand for our AI services and in a future release, we may introduce some transparency into training waiting times when you submit a solution for training. Usually, we see spikes in demand for the AI Services at the beginning of the month, quarter and year, when solutions are scheduled to be retrained. Therefore, we recommend that you train the Predictive Intelligence solutions ahead of time in case there is high demand and avoid any tight schedules when deploying solutions that still require training.