Open YuviGold opened 3 years ago
Thanks for your report.
Ideally this data could be provided via copy.Options.Progress
, not (or at least not primarily) as a plain text in a log stream.
That channel does currently provide “offset” + blob ID, so a reader maintaining its own timestamp can calculate instantaneous speed; but it does not provide metadata like list of layers to copy and expected size that would allow to make a prediction of remaining time.
Currently, in case anyone would want to calculate image pull speed, he would have to
But this is the best-case scenario. It doesn't take into account caching, failures, compression, etc.
I suggest adding a new feature for notifying about the download rate of each layer to the user. Either by a new operation(which requires saving those statistics to be gathered later on) or by real-time debug logs of the pull operation.
/cc @vrothberg