Related to #64: long-term, it may be beneficial for us to have the dark archive set up exactly how we'd need it for re-ingesting all batches. This is a very low-priority task because of the time involved in adjusting archived data compared to the value it brings us. There's a fair amount of work to mitigate the risk of breaking data, and all this does is save a fairly small amount of time in the event of a catastrophic data-loss event. That said, doing this means a more consistent archive state and a bit of storage savings, as well as a less error-prone DR process.
Related to #64: long-term, it may be beneficial for us to have the dark archive set up exactly how we'd need it for re-ingesting all batches. This is a very low-priority task because of the time involved in adjusting archived data compared to the value it brings us. There's a fair amount of work to mitigate the risk of breaking data, and all this does is save a fairly small amount of time in the event of a catastrophic data-loss event. That said, doing this means a more consistent archive state and a bit of storage savings, as well as a less error-prone DR process.