I like the sound of it.
A useful first step (IMHO) will be to retry the write requests at the highest possible level.
From there we'll know we have a safety net and can try to refine this at a finer-grained level.
I like the sound of it.
A useful first step (IMHO) will be to retry the write requests at the highest possible level.
From there we'll know we have a safety net and can try to refine this at a finer-grained level.