Typical raw (as sampled, not resampled) time series data does not have regular timestamps. The origin of this could be having multiple machines, sampling rates, kepware instances or sample compression (drop uninteresting samples) in Historians like OSI PI.
Typically also, not every variable has a sample for every timestamp, e.g. analogs are sampled quicker than digitals or different schedules in manual/automatic sampling.
Can TWA please support the ingestion of ths type of data?