[BUG] High sustained CPU usage (~100% per instance) even when idle
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
Description
Two instances of Claude Code CLI running simultaneously each consume ~100% CPU sustained, even between interactions when both should be idle. On a 4-core machine this makes the system noticeably sluggish.
Environment
- Claude Code version: 2.1.29
- Node.js: v22.19.0
- OS: Linux Mint 21.3 (kernel 5.15.0-164-generic)
- Hardware: 4 CPU cores, 7.5 GB RAM
- Terminal: Guake (but issue is in the claude process itself)
Observed behavior
Both claude processes pin at ~100% CPU each continuously, even when no tool calls are running and no output is being streamed:
PID USER %CPU %MEM RES COMMAND
3225 jan 102% 9.6% 756 MB claude
5342 jan 89% 8.4% 660 MB claude
A second measurement 2 minutes later with both sessions idle between interactions:
PID USER %CPU %MEM RES COMMAND
3225 jan 101% 9.9% 781 MB claude
5342 jan 97% 9.0% 705 MB claude
Each process has ~13-14 threads. Virtual memory mapped is ~71 GB per process (normal for V8 but noted for completeness).
---
What Should Happen?
Expected behavior
Between interactions, CPU usage should drop to near zero since the CLI is mostly waiting for user input or network responses.
Error Messages/Logs
Steps to Reproduce
- Open two terminal tabs
- Run claude in each
- Have a conversation in one; leave the other idle
- Observe CPU usage with top
Claude Model
Opus
Is this a regression?
Yes, this worked in a previous version
Last Working Version
_No response_
Claude Code Version
2.1.29
Platform
Anthropic API
Operating System
Ubuntu/Debian Linux
Terminal/Shell
Other
Additional Information
- No MCP servers configured
- The high CPU is in the claude Node.js process itself, not in any child/tool processes
- Single instance also shows ~100% CPU, it's not contention between the two
This issue has 12 comments on GitHub. Read the full discussion on GitHub ↗