Skip to content

Fix excessive memory allocation for sched_times (#320) #389

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
florianl merged 1 commit into from
Mar 19, 2025
Merged

Fix excessive memory allocation for sched_times (#320) #389

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 8 additions & 10 deletions tracer/tracer.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -535,26 +535,24 @@ func loadAllMaps(coll *cebpf.CollectionSpec, cfg *Config,
adaption := make(map[string]uint32, 4)

const (
// The following sizes X are used as 2^X, and determined empirically

// The following sizes X are used as 2^X, and determined empirically.
// 1 million executable pages / 4GB of executable address space
pidPageMappingInfoSize = 20

pidPageMappingInfoSize = 20
stackDeltaPageToInfoSize = 16
exeIDToStackDeltasSize = 16
)

adaption["pid_page_to_mapping_info"] =
1 << uint32(pidPageMappingInfoSize+cfg.MapScaleFactor)

adaption["stack_delta_page_to_info"] =
1 << uint32(stackDeltaPageToInfoSize+cfg.MapScaleFactor)

// To not loose too many scheduling events but also not oversize
// sched_times, calculate a size based on some assumptions.
// On modern systems /proc/sys/kernel/pid_max defaults to 4194304.
// Try to fit this PID space scaled down with cfg.OffCPUThreshold into
// this map.
adaption["sched_times"] = (4194304 * cfg.OffCPUThreshold) / support.OffCPUThresholdMax
// To not lose too many scheduling events but also not oversize sched_times,
// calculate a size based on an assumed upper bound of scheduler events per
// second (1000hz) multiplied by an average time a task remains off CPU (3s),
// scaled by the probability of capturing a trace.
adaption["sched_times"] = (4096 * cfg.OffCPUThreshold) / support.OffCPUThresholdMax
Comment on lines +553 to +555
Copy link
Copy Markdown
Contributor

@rockdaboot rockdaboot Mar 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not a blocker, but it is not obvious to me how 4096 reflects the mentioned "average time a task remains off CPU (3s)". How exactly has the number 4096 been determined and is this the the best value or does it need further refinement?

Copy link
Copy Markdown
Member Author

@christos68k christos68k Mar 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pow2(1000*3)


for i := support.StackDeltaBucketSmallest; i <= support.StackDeltaBucketLargest; i++ {
mapName := fmt.Sprintf("exe_id_to_%d_stack_deltas", i)
Expand Down