[PR #421] [MERGED] feat: add scheduler throughput benchmarks #423

New issue

Closed

opened 2026-10-29 14:23:16 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-29 14:23:16 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/kvcache-ai/AgentENV/pull/421
Author: @GxmzG
Created: 10/22/2026
Status: ✅ Merged
Merged: 10/28/2026
Merged by: @yingdi-shan

Base: main ← Head: feat/benchmark-scheduler-throughput

📝 Commits (1)

49aedd1 feat: add scheduler throughput benchmarks

📊 Changes

2 files changed (+153 additions, -1 deletions)

View changed files

📝 services/README.md (+15 -1)
➕ services/scheduler/internal/service_benchmark_test.go (+138 -0)

📄 Description

Summary

Add scheduler throughput benchmarks for direct Schedule calls and in-memory gRPC calls, with 1 or 32 nodes and 1, 2, 4, or 8 workers.

Measurement

Report successful calls in schedule/s. Setup is excluded; the gRPC result includes handler and serialization work, but neither result includes TCP traffic or sandbox startup.

Validation

Service tests, format check, vet, and race check passed. Three benchmark runs completed;，and at 32 nodes and 4 workers, direct calls measured 353,658–355,714 schedule/s and in-memory gRPC measured 81,693–84,433 schedule/s.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/kvcache-ai/AgentENV/pull/421 **Author:** [@GxmzG](https://github.com/GxmzG) **Created:** 10/22/2026 **Status:** ✅ Merged **Merged:** 10/28/2026 **Merged by:** [@yingdi-shan](https://github.com/yingdi-shan) **Base:** `main` ← **Head:** `feat/benchmark-scheduler-throughput` --- ### 📝 Commits (1) - [`49aedd1`](https://github.com/kvcache-ai/AgentENV/commit/49aedd19ebd9e60c80014b549456cb9ce2371ab0) feat: add scheduler throughput benchmarks ### 📊 Changes **2 files changed** (+153 additions, -1 deletions) <details> <summary>View changed files</summary> 📝 `services/README.md` (+15 -1) ➕ `services/scheduler/internal/service_benchmark_test.go` (+138 -0) </details> ### 📄 Description ## Summary Add scheduler throughput benchmarks for direct `Schedule` calls and in-memory gRPC calls, with 1 or 32 nodes and 1, 2, 4, or 8 workers. ## Measurement Report successful calls in `schedule/s`. Setup is excluded; the gRPC result includes handler and serialization work, but neither result includes TCP traffic or sandbox startup. ## Validation Service tests, format check, vet, and race check passed. Three benchmark runs completed;，and at 32 nodes and 4 workers, direct calls measured 353,658–355,714 schedule/s and in-memory gRPC measured 81,693–84,433 schedule/s. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-29 14:23:16 +00:00