We are observing large upload delays of more than 30 seconds, i.e. the data become visible in e.g. FleetManager only after this delay. This is too long for our partner to react on critical events. We are uploading data to MindSphere via MindConnect Integration on Premise. Is this delay really unavailable?
The visibility of data in fleet manager is dependent on the availability of the aggregates which are built asynchronously when the data is uploaded.
This doesn't necessarily mean that the raw data can't be queried via the IoT TimeSeries API.
Due to the asynchronous nature of the mindsphere (which is absolutely required to be able to scale properly) it can't be guaranteed that the data will always be there in < 30 sec. This looks to me like a use case which is better suited for edge analytics where e.g. training of the algorithm is done in the cloud, but the processing is executed on the edge.