[OneDev #24] fix(perf): resolve performance issues — emergency #15

Closed
opened 2026-05-24 04:46:37 +00:00 by joseph · 1 comment
Owner

Imported from OneDev issue #24 (id 35504)
Original project: internal-joseph
Original state: Closed
Original submitterId: 5
Original submitDate: 2026-04-29T14:02:11.517+00:00
Original lastActivity: {'date': '2026-04-29T15:12:35.564+00:00', 'description': "changed state to 'Closed'", 'userId': 5}

Original fields:

  • Type: New Feature
  • Priority: Normal
  • Assignees: None

Emergency-status performance issues need to be diagnosed and resolved. Capture symptoms, scope (which host/service/app), and reproduction; then triage root cause and ship the fix.

Why

Flagged as emergency — blocking work or degrading a live service. Velocity matters; details to be filled in once the affected surface is named.

Open questions

  • Which system is degraded? (workstation, specific LXC/VMID, web service, model inference, etc.)
  • Symptom shape: latency, throughput, hangs, OOM, CPU pegged?
  • When did it start, and what changed around that time?
Imported from OneDev issue #24 (id 35504) Original project: internal-joseph Original state: Closed Original submitterId: 5 Original submitDate: 2026-04-29T14:02:11.517+00:00 Original lastActivity: {'date': '2026-04-29T15:12:35.564+00:00', 'description': "changed state to 'Closed'", 'userId': 5} Original fields: - Type: New Feature - Priority: Normal - Assignees: None Emergency-status performance issues need to be diagnosed and resolved. Capture symptoms, scope (which host/service/app), and reproduction; then triage root cause and ship the fix. ## Why Flagged as emergency — blocking work or degrading a live service. Velocity matters; details to be filled in once the affected surface is named. ## Open questions - Which system is degraded? (workstation, specific LXC/VMID, web service, model inference, etc.) - Symptom shape: latency, throughput, hangs, OOM, CPU pegged? - When did it start, and what changed around that time?
Author
Owner

Imported from OneDev comment id 523 on issue #24
Original userId: 5
Original date: 2026-04-29T15:12:34.564+00:00

Resolved live in chat 2026-04-29. Root cause: 1,151 orphaned pwsh.exe/conhost.exe processes left behind by exited Claude Code sessions, accumulating ~18 GB commit charge on a 31 GB workstation and triggering a paging storm. Mitigation deployed:

  • One-shot sweep killed orphans, dropped commit from 103 GB → 70 GB
  • Scheduled task ReapOrphanShells runs every 15 min on pct-joseph-nu, kills pwsh/conhost where parent PID is dead and age >2h
  • Script: C:\Users\Joseph\.claude\scripts\reap-orphan-shells.ps1
  • Memory note saved at system_orphan_shell_reaper.md

Closing — issue was filed mistakenly during diagnosis.

Imported from OneDev comment id 523 on issue #24 Original userId: 5 Original date: 2026-04-29T15:12:34.564+00:00 Resolved live in chat 2026-04-29. Root cause: 1,151 orphaned pwsh.exe/conhost.exe processes left behind by exited Claude Code sessions, accumulating ~18 GB commit charge on a 31 GB workstation and triggering a paging storm. Mitigation deployed: - One-shot sweep killed orphans, dropped commit from 103 GB → 70 GB - Scheduled task `ReapOrphanShells` runs every 15 min on pct-joseph-nu, kills pwsh/conhost where parent PID is dead and age >2h - Script: `C:\Users\Joseph\.claude\scripts\reap-orphan-shells.ps1` - Memory note saved at `system_orphan_shell_reaper.md` Closing — issue was filed mistakenly during diagnosis.
Sign in to join this conversation.
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
joseph/internal-joseph#15
No description provided.