Add workload-level profiling (draft — provisory) - #1054
Draft
artursarlo wants to merge 1 commit into
Draft
artursarlo wants to merge 1 commit into
artursarlo wants to merge 1 commit into
Conversation
Sync Pinterest's dynamic-profiling subsystem forward onto upstream as a coherent, self-consistent unit so the workload-level profiling feature builds and runs against intel/gprofiler master. This brings the workload inventory feature together with the supporting subsystems it depends on at runtime: - Workload-level profiling: heartbeat workload inventory collector (namespace/pod/container/process discovery, configurable workload name/kind label keys), spec, and fast spec tests. - Heartbeat / dynamic profiling: queue-based command control, continuous and ad-hoc profiling slots, heartbeat perf limits as CLI args. - PMU: multi-PMU perf event detection and manager. - Metrics: metrics publisher for heartbeat error-budget/SLI reporting. - mTLS: mutual TLS support for the agent heartbeat client. The dynamic-profiling core hard-depends on the PMU and metrics modules, so they are included to keep the import graph consistent with the fork's tested code. requirements.txt keeps upstream's cpuid pin and adds bitmath (used by the memory manager). Verified: full-package import sweep passes and the workload fast tests (tests_fast) pass. Co-authored-by: Lucas <lpenhademoura@pinterest.com> Co-authored-by: ashokchatharajupalli <achatharajupalli@pinterest.com> Co-authored-by: prashantpatel <prashantpatel@pinterest.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Brings Pinterest's workload-level profiling feature onto upstream as a single, self-consistent commit.
The workload code is entangled with the PMU, metrics-publisher, and mTLS work at the import/runtime level, so this syncs the whole dynamic-profiling subsystem forward as a coherent unit (rather than a partial slice that wouldn't build):
requirements.txtkeeps upstream'scpuidpin and addsbitmath(used by the memory manager).Verified: full-package import sweep passes (0 failures); workload fast tests (
tests_fast) pass (50).Co-authored with Pinterest contributors (see commit trailers).