In #9 we implemented the means to record failures to disk.
Although this allows calculated data to be recorded, we don't currently provide a means to replay those files in order to try and write them into the database again.
One of the consequences of moving to batched writes in #15 is that we're no longer able to record the information that would be necessary for a replay.
This means that this issue needs closing as Won't Fix, but that's probably OK, the idea was never to become a durable queue so whilst it would have been nice to have it's not essential.
Activity
05-Feb-23 23:49
assigned to @btasker
07-Feb-23 19:56
One of the consequences of moving to batched writes in #15 is that we're no longer able to record the information that would be necessary for a replay.
This means that this issue needs closing as
Won't Fix, but that's probably OK, the idea was never to become a durable queue so whilst it would have been nice to have it's not essential.