[Bug] Orphaned subagent processes not reaped on connection drop; memory leaks to ~8 GB per resume cycle

Status Closed — not planned
Reported on v2.1.193
Maintainer reply None cached
Activity 2 comments · opened Jun 27, 2026 · closed Aug 13, 2026

Bug Description
Multi-agent Workflow subagents are orphaned (not reaped) when the connection drops mid-run; each resume stacks new ones, and orphaned processes balloon to ~8 GB RSS each.

Environment Info

  • Platform: darwin
  • Terminal: zed
  • Version: 2.1.193
  • Feedback ID: 40c41108-0d87-46dc-a2d2-9dc532950f1c

Errors

[]

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗