Hello, On Mon, Jan 13, 2025 at 10:39:02AM -0800, Shakeel Butt wrote: ... > Here I think the supportive data we can show is the (1) non-memory stats > readers not needing to spend time on memory stats flushing and (2) with > per-subsystem update tree, have we increased the cost of update tree > insertion in general? > > Anything else you think will be needed? Anything that shows: 1. It doesn't cause noticeable regressions for generic use cases. 2. It has noticeable benefits in the targeted use cases. The above is a bit willy-nilly but 2 should be fairly straightforward. For 1, the main worry would be whether high(er) frequency monitoring using something like systemd-cgtop or below becomes noticeably more expensive on a large and busy system. Thanks. -- tejun
