Debug Slow Mac

Systematically diagnose macOS performance issues by checking CPU, memory, and identifying the worst offenders.

Workflow

1. Get System Overview

Run these in parallel to get a quick snapshot:

# Load average and basic stats
uptime

# Memory pressure
vm_stat | head -10

# Top CPU consumers (snapshot)
ps -Ao pid,pcpu,pmem,comm -r | head -15

Key metrics to report:

• •Load average: Should be roughly equal to CPU core count. Much higher = overloaded
• •Pages free vs wired: Low free + high wired = memory pressure
• •CPU idle %: Below 20% idle = CPU bottleneck

2. Identify Top Offenders

Get detailed process info sorted by CPU:

ps -Ao pid,pcpu,pmem,comm -r | head -20

For processes with high CPU, get their full command line:

ps -eo pid,pcpu,pmem,command -r | head -15

3. Check Docker/OrbStack (if running)

docker stats --no-stream --format "table {{.Name}}\t{{.CPUPerc}}\t{{.MemUsage}}" 2>/dev/null | head -15

4. Map High-CPU Processes to Tmux Sessions

When processes like claude, nvim, node, etc. are consuming resources, identify exactly which tmux session/window/pane they're in.

Step 1: Get all tmux panes with their TTYs

tmux list-panes -a -F '#{session_name}:#{window_index}.#{pane_index} #{pane_tty} #{pane_current_command}'

Step 2: For a specific high-CPU PID, trace to its tmux pane

# Get the TTY for the process
ps -o pid,tty,command -p <PID>

# Match the TTY (e.g., ttys043) against tmux panes
tmux list-panes -a -F '#{session_name}:#{window_index}.#{pane_index} #{pane_tty}' | grep ttys043

Step 3: For processes that spawn children (like claude spawning node)

Trace the parent chain to find the original TTY:

# Get parent PID
ps -o pid,ppid,tty,command -p <PID>

# If TTY shows "??", check the parent process
ps -o pid,ppid,tty,command -p <PPID>

# Keep tracing up until you find a real TTY

Step 4: Build a summary of which tmux locations have high-CPU processes

# For each high-CPU process, find its tmux location
for pid in <PID1> <PID2> <PID3>; do
  tty=$(ps -o tty= -p $pid 2>/dev/null | tr -d ' ')
  if [ "$tty" != "??" ] && [ -n "$tty" ]; then
    pane=$(tmux list-panes -a -F '#{session_name}:#{window_index}.#{pane_index} #{pane_tty}' | grep "$tty" | cut -d' ' -f1)
    cmd=$(ps -o comm= -p $pid)
    echo "$pane -> $cmd (PID $pid)"
  fi
done

Example Output to Provide User

Present findings like:

Always provide the tmux switch-client -t <location> command so the user can jump directly to the offending pane.

5. Check for Common Culprits

Node.js / TypeScript LSP

ps -eo pid,pcpu,pmem,command | grep -E 'tsserver|vtsls|node.*typescript' | grep -v grep

High CPU here usually means:

• •Large TypeScript project being type-checked
• •Multiple neovim/VSCode instances with LSP

Electron Apps

ps -eo pid,pcpu,pmem,command | grep -E 'Electron|Helper.*Renderer' | grep -v grep

Common offenders: Slack, Discord, Linear, VSCode, browsers

Claude Code instances

ps -eo pid,pcpu,pmem,command | grep -E '[c]laude' | head -10

6. Provide Actionable Recommendations

Based on findings, suggest:

1. •For Docker/OrbStack: orbctl stop if not needed
2. •For specific containers: docker stop <name>
3. •For Storybook/dev servers: pkill -f storybook or pkill -f vite
4. •For runaway Node processes: Identify and kill specific PIDs
5. •For Electron apps: Quit via GUI or pkill -f <AppName>
6. •For LSP issues: Close editor buffers or restart editor

Always offer to kill specific processes and provide the exact command.

Output Format

Present findings in a clear table:

Then provide:

1. •Root cause summary
2. •Specific commands to fix
3. •Ask if user wants you to run the fix commands