Skip to content

Add pprof endpoint flag to the operator - #2805

Open
rajathagasthya wants to merge 1 commit into
NVIDIA:mainfrom
rajathagasthya:pprof-endpoint
Open

Add pprof endpoint flag to the operator#2805
rajathagasthya wants to merge 1 commit into
NVIDIA:mainfrom
rajathagasthya:pprof-endpoint

Conversation

@rajathagasthya

Copy link
Copy Markdown
Contributor

Description

Add a --pprof-bind-address flag wired to controller-runtime's built-in
pprof server (ctrl.Options.PprofBindAddress), for profiling the
operator's memory and CPU at large node counts. The endpoint is disabled
by default and only serves when an address (e.g. ":6060") is given.

The address is exposed as the operator.pprof.bindAddress chart value,
passed to the operator only when set.

Checklist

  • No secrets, sensitive information, or unrelated changes
  • Lint checks passing (make lint)
  • Generated assets in-sync (make validate-generated-assets)
  • Go mod artifacts in-sync (make validate-modules)
  • Test cases are added for new code paths

Testing

go build ./cmd/gpu-operator; helm template with and without
operator.pprof.bindAddress set (flag rendered only when set); helm lint.

@rajathagasthya rajathagasthya self-assigned this Aug 26, 2026
Add a --pprof-bind-address flag wired to controller-runtime's built-in
pprof server (ctrl.Options.PprofBindAddress), for profiling the
operator's memory and CPU at large node counts. The endpoint is
disabled by default and only serves when an address (e.g. ":6060")
is given.

Expose the address as the operator.pprof.bindAddress chart value,
passed to the operator only when set.

Signed-off-by: Rajath Agasthya <ragasthya@nvidia.com>
@coderabbitai

coderabbitai Bot commented Aug 31, 2026

Copy link
Copy Markdown

Review Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: QUIET

Plan: Enterprise

Run ID: d38125fa-caf9-4cb5-a330-ec9d681378b4

📥 Commits

Reviewing files that changed from the base of the PR and between fa49ea6 and 41c6c1a.

📒 Files selected for processing (3)
  • cmd/gpu-operator/main.go
  • deployments/gpu-operator/templates/operator.yaml
  • deployments/gpu-operator/values.yaml

Included review availability: Your plan provides up to 12 included reviews per hour; 11 remain after this review.


📝 Walkthrough

Walkthrough

The operator adds a --pprof-bind-address command-line option. The configured value is passed to the controller-runtime manager through PprofBindAddress. The Helm chart adds an operator pprof.bindAddress value and emits the command-line argument when the value is set. The default empty value keeps the profiling endpoint disabled.

Merge Risk: 🔵 Low · up to 41c6c

The PR adds an optional pprof endpoint that is disabled by default, but enabling it on a broad address such as :6060 could expose profiling data to unintended clients. The change is mergeable with owner awareness that access protection or a localhost-only binding must be confirmed.


Comment @coderabbitai help to get the list of available commands.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants