@Diederik I tried to replicate it using the attached node-red flow, which basically sends sends a message every 0.1 sends to one of 40 tags at random. These tags vary in length from "ns_3_s_Area_3_Cooling_Pump_2_Status" to "_ns_3_s_Area_4_Boiler_3_Combustion_Efficiency_Optimization_Module_Carbon_Monoxide_Emission_Reduction_Target_Achievement" (I asked ChatGPT to generate some random tags) and some of them begin or end with with underscores. So far i only achieved to make the frontend a lag a bit. Is there something else i might have missed with my dataset ?