Improve scheduling of the memory profiler. (#35618)

Switch the delay to be used between the end of a previous run and the next, instead of the start
of consecutive runs. That ensure that we don't enqueue messages when processing is slower than
the delay.

Signed-off-by: webbeef <me@webbeef.org>
This commit is contained in:
webbeef 2025-02-23 12:04:57 -08:00 committed by GitHub
parent d446b63cfd
commit 678f0c9fdc
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 25 additions and 2 deletions

1
Cargo.lock generated
View file

@ -5811,6 +5811,7 @@ dependencies = [
"base",
"ipc-channel",
"libc",
"parking_lot",
"profile_traits",
"regex",
"serde",