Skip to content

Warm Pool Feasibility Study [Spec + Experiments]#2199

Closed
rhuss wants to merge 2 commits into
NVIDIA:mainfrom
rhuss:6111-warm-pool-feasibility
Closed

Warm Pool Feasibility Study [Spec + Experiments]#2199
rhuss wants to merge 2 commits into
NVIDIA:mainfrom
rhuss:6111-warm-pool-feasibility

Conversation

@rhuss

@rhuss rhuss commented Jul 9, 2026

Copy link
Copy Markdown

Summary

Feasibility study for warm-pooling sandbox pods to reduce startup latency from 8-12s to under 2s. Adds experiment scripts, K8s manifests, a readiness sidecar, and an RFC scaffold for recording results.

  • 6 measurement scripts covering cold-start baseline, warm pool claims, readiness gates, sidecar readiness, env var injection, and combined best-case scenarios
  • Shared library functions for CSV output, statistics, timestamps, and pod event collection
  • K8s manifests for SandboxTemplate, SandboxWarmPool, SandboxClaim, image prepull DaemonSet
  • Go readiness sidecar with signal-file-based health endpoint
  • RFC template structured for latency data, configuration comparison, and architecture recommendations

Related Issue

#2157

Testing

  • All shell scripts pass bash -n syntax check
  • Go sidecar builds with CGO_ENABLED=0
  • Deep review completed by 5 specialized agents (correctness, architecture, security, production readiness, test quality)
  • 7 findings fixed across 2 fix rounds (see specs/001-warm-pool-feasibility/review-findings.md)

Checklist

  • Scripts produce CSV output with per-phase timestamps (FR-004, FR-005)
  • Manifests cover SandboxTemplate, SandboxWarmPool, SandboxClaim (FR-006)
  • Readiness gate pod manifest with custom condition (FR-007)
  • Sidecar readiness container with controllable endpoint (FR-008)
  • RFC scaffold with all required sections (FR-009, FR-010)
  • Deep review findings documented

Assisted-By: 🤖 Claude Code

rhuss added 2 commits July 9, 2026 14:33
Brainstorm, specification, implementation plan, task breakdown, and
review guide for evaluating Agent Sandbox warm pooling on OpenShift.
Measures whether sandbox startup latency can be reduced from 8-12s
to under 2s using SandboxWarmPool/SandboxClaim extension CRDs.

Assisted-By: 🤖 Claude Code
Add measurement scripts for cold-start baseline, warm pool claims,
readiness gate patterns, sidecar readiness, env var injection, and
combined best-case scenarios. Each script produces CSV output with
per-phase timestamps for latency analysis.

Includes:
- 6 measurement scripts with shared library functions
- K8s manifests for SandboxTemplate, WarmPool, Claims, and DaemonSet
- Go readiness sidecar with signal-file-based health endpoint
- RFC template document with structured sections for results
- Deep review findings from 5-agent code review

Assisted-By: 🤖 Claude Code
Signed-off-by: Roland Huß <rhuss@redhat.com>
@rhuss rhuss requested review from a team, derekwaynecarr, maxamillion and mrunalp as code owners July 9, 2026 14:35
@copy-pr-bot

copy-pr-bot Bot commented Jul 9, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@rhuss rhuss closed this Jul 9, 2026
@rhuss

rhuss commented Jul 9, 2026

Copy link
Copy Markdown
Author

sorry, opened by accident.

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.

1 participant