Hi there,
Currently we are using "normal" indexes instead of data streams in our application. But in the last days I was faced to ILM and it could be easier to implement with data streams. So I had a look if data streams could be used. But for the following question I could not find any answer or detailed concept descriptions:
Is it possible to "backfill" a data stream?
In our application we have e.g. events with a timestamp and detailed data. So in general we have a timestamp field which could be used for data stream. But it could happen that the export of the events fails from time to time and is repeated afterwards manually. So the timestamps are older than the latest entry in the data stream. What would happen then? Is the new data sorted in somehow or just appended to the end of that data stream index? What would happen if the timestamp is in a range of data which is already in warm or cold phase?
Thank you for your support