qwen-code issue: idle Managed Runtime worker holds 91 MiB, mostly never-run modules
wenshao · ghdev · 2026-11-04
A detailed performance issue on QwenLM/qwen-code dissects why every resident sandbox's Managed Runtime worker idles at 91 MiB of private memory versus 7.7 MiB for a bare Node process.
- Most of the gap is source and metadata of modules loaded but never executed; a Node process that merely imports core's Config idles at 87–92 MiB.
- After forced GC, the V8 heap still holds 63 MiB, but bytecode is only 2.6 MiB and compiled code 2.8 MiB — the floor is the loaded modules themselves.
- The worker entry's static import closure spans 2,046 modules / 12 MiB, with Config alone accounting for 1,576 modules / 8.4 MiB; express→body-parser (1.4 MiB, including two copies of iconv-lite) and google-auth-library are pulled in for single constants or minor features.
Includes a dominator analysis table from the esbuild metafile — a solid case study in Node module-level memory profiling.
More from coding & agent
- Dev proposes LLM-generated ultra-rare content to build anime-style MMORPGs — rickasaurus · 2026-11-04
- Dev builds Chrome extension to bring back the classic Twitter bird on X — joonasvirtanen · 2026-11-04
- Developer lets his app record 'wishes' and auto-spawn OpenCode background agents to fulfill them — SubtleGradient · 2026-11-04
- Post-train your model on your own agent harness with OpenHands, Harbor, SkyRL and Snowflake — StasBekman · 2026-11-04
- Google Antigravity redesign leaked with context usage and quota indicators — clmt · 2026-11-04
- Microsoft's 4B Mage-VL-Cowork Hits 66% on OSWorld-Verified, Near Human Level — microsoft · 2026-11-04