#Performance issues on migrating from v0.11.7 to v0.16.2
1 messages · Page 1 of 1 (latest)
👋 heya! yeah, oof, that's a bit of a jump, hope aside from the performance issues that it's going okay?
out of curiosity, when you say "50% disk", what are you measuring? is this the amount of disk usage on the node?
i wonder if there's still disk contention happening, where the bandwidth of the disk is actually getting saturated.
i think you would measure this using iostat command, you're probably interested in the iowait and r_await and w_await columns
one pretty piece of low hanging fruit to verify - i would make sure that the /var/lib/dagger path in the dagger engine container is mapped to a persistent volume. leaving it as-is in a simple setup means that it gets mapped to an overlayfs which is slow