CPU Pressure stalls caused by Tokio in the storagenode application (Linux)

If your node is showing a lot of lost upload/download races: that’s definately something that could be looked into. Are you measuring with the successrate.sh script? (I know it’s old, I just don’t know if there’s a better util these days)

I think “normal” is still 90%+ for most people. But I have seen people mention going lower sometimes.