MC-4455 — Observability hardening: stuck-ticket alerter + orphan tmux sweep

Home Board MC-4455

Observability hardening: stuck-ticket alerter + orphan tmux sweep

**Finding 1:** get_stuck_tickets() is pull-only — waits for human to ask. No active poller. **Finding 2 (Kimi):** No periodic tmux ls vs runtime_sessions diff. Orphaned CLI proc...

State Done Next Action Closed Owner Luci Runtime Closed Age 14d ago

← MC-4455

Ticket is done; runtime is closed. · profile claude_opus_1m_medium · cwd /home/lucienne/workspace/.claude/worktrees/pool-2 · uptime 13d 20h · last activity 13d 18h ago

Description

MC-4455

**Finding 1:** get_stuck_tickets() is pull-only — waits for human to ask. No active poller. **Finding 2 (Kimi):** No periodic tmux ls vs runtime_sessions diff. Orphaned CLI processes accumulate silently, leaking cost. **Action:** 1. Add scheduled stuck-ticket poller (every N minutes, Telegram alert) 2. Add orphan sweep: `tmux ls` vs `runtime_sessions` diff, kill+log orphans 3. Consider combining into a single health-check daemon **Source:** FINAL-ASSESSMENT.md Items #5 + #6

Activity

done

Luci is working...

Details — Done · Medium · Luci ▼

State

Done

Closed

State

Priority

People

Owner (assigned to)

Controller

Timing / Details

Due Date

Snooze

Source api (human)

Project Life

Created 14d ago

Updated 14d ago

Advanced / Operator evidence

Routing owner

Routes via

Operator console

Evidence

Ticket is done; runtime is closed. ticket_runtime_done profile claude_opus_1m_medium · cwd /home/lucienne/workspace/.claude/worktrees/pool-2 · uptime 13d 20h · last activity 13d 18h ago MC is visibility-only. Hermes Luci launches and gates work outside MC, then mirrors evidence/status here. Raw console: luci · claude/anthropic/opus[1m]

Workflow

Start Dev Review + QA ▾

Select phases to include:

Research (scott) Implement (larry) (required) Council Review (council) Code Review (luci) Validate (tessa) (required) Sign-off (atlas)

Agents

Review Gates

Decision

WAT routing: choose an agent, review gate, or decision. Buttons use the live runtime when one is attached.

triage-bot 14d ago

auto-triage: assigned to **luci** (confidence 1.00); moved to `todo` for pickup -- skill: `coding, scheduling, infrastructure, system health` > The ticket involves adding scheduled tasks, scripting system health checks, and managing processes on a Linux server, which are core responsibilities for luci.

luci 14d ago

Luci picking up this ticket (background worker).

luci 14d ago

Ticket picked up by Luci via MC dispatcher. MC-4455: Observability hardening: stuck-ticket alerter + orphan tmux sweep Work this ticket in the live tmux runtime. Use DONE:, REVIEW:, or QUESTION: when you need MC to reflect the next state.

Luci 14d ago

Mapped MC-4455 infrastructure: get_stuck_tickets, runtime_sessions schema, tmux naming, notify.py, scheduler format, queue_reaper, and existing orphan cleanup.

luci 14d ago

Workflow action `review_ready` applied. Work ready for review.

api 14d ago

Orchestrator gate: CLOSING as already-implemented. Both action items already ship live in the observability-sweep scheduler task (tasks/observability-sweep.md, scripts/observability_sweep.py, 8.4K), enabled, every 15 min, parent MC-3365/MC-3370: (1) Finding 1 stuck-ticket alerter — diffs /api/v1/tickets/stuck vs last-seen set, Telegram-alerts only NEW entries with 1h dedup, closing the pull-only gap in get_stuck_tickets(); (2) Finding 2 orphan tmux sweep — diffs 'tmux ls' (mc-MC-* sessions) vs runtime_sessions WHERE status IN (starting,running,idle,needs_input), flags leaked CLI/browser procs + stale db rows, 6h dedup. Verified healthy: task_runs completed at 13:01/13:15/13:30/13:45 today. The worker correctly mapped the infra and found the existing cleanup — REVIEW was the right call; no new code needed. FINAL-ASSESSMENT items #5+#6 reconciled to MC-3365/3370. No duplicate task created (key rule #6).

luci 14d ago

Shadow reviewer verdict: FAIL (advisory only). Ticket was gated to done by the orchestrator/operator — not reopening; the gate is authoritative. Reasons: The committed diff contains only changes to a graph report markdown file (graphify-out/GRAPH_REPORT.md) updating community lists and metadata. It contains absolutely no code changes related to the ticket requirements: no scheduled stuck-ticket poller, no orphan tmux sweep logic, no health-check daemon, and no tests. The worker's claim that these features 'already ship live' in observability_sweep.py is unsupported by the provided diff, which shows no modifications to any Python scripts, scheduling tasks, or observability infrastructure. Gaps: - No implementation of a scheduled stuck-ticket poller (Finding 1) - No implementation of an orphan tmux sweep diffing tmux ls vs runtime_sessions (Finding 2) - No health-check daemon or combined scheduler task added - No test evidence or smoke tests for the claimed functionality - The diff is entirely unrelated to the ticket (only a graph report metadata update)

Live ▼

No activity yet

←