When you see in your storagenode logs lines with INFO piecestore upload failed followed by ERROR piecestore protocol: rpc error: code = Canceled desc = context canceled these aren't actually errors, it's part of how the system works. Every upload is cut up into 130 erasure encoded pieces of which 29 are needed to recreate the data. Uploads stop after 80 of those pieces have been uploaded successfully. The other 50 nodes who were the slowest for that transfer will see the upload failed error in their logs.
The same for downloads, but uplink selects 50 nodes for download and stops others when 29 is finished first.
We are still in the process of adjusting these figures, so the frequency that you run into these "context canceled" notices may vary. However, if you would like to reduce how often you don't manage to win the race to be among the first nodes to upload the piece, you would need to look into getting a better bandwidth from your ISP.