Skip to content

Latest commit

 

History

History
49 lines (39 loc) · 2.09 KB

File metadata and controls

49 lines (39 loc) · 2.09 KB

Triage Guide

This guide standardizes severity labels and SLAs for Plan Manager issues.

Labels

  • type:bug — defect in behavior
  • type:enhancement — improvement or polish
  • type:docs — documentation only
  • area:* — component or subsystem (e.g., area:tools, area:server, area:docs)
  • needs-info — missing reproduction or details \n+Performance topics: use area:performance in addition to type:enhancement unless there is a clear functional defect (then use type:bug).

Severity (Priority)

  • P0 — Crash/data loss/security; blocks primary flows.
  • P1 — Broken but with workaround; major UX or correctness impact.
  • P2 — Minor issue/polish; does not block shipping. \n+Examples:
  • P1: Feature works but is unacceptably slow for typical inputs; or incorrect output with readily available workaround.
  • P2: Minor latency regression in non-critical path; cosmetic misalignment.

SLAs (targets)

  • P0 — Acknowledge: 24h; Fix: 72h or hotfix plan.
  • P1 — Acknowledge: 48h; Plan/Fix: within one iteration.
  • P2 — Batch and address during polish windows.

Triage Routine

  • Initial: apply labels, set P0/P1/P2, add repro steps or mark needs-info.
  • Weekly: review new/untriaged items; re-evaluate severities; assign owners.

Weekly cadence & ownership

  • Schedule: every Monday, 10:00 UTC (adjust as needed for your team).
  • Participants: maintainer + last week’s contributors.
  • Reminders: calendar event and chat reminder 15 minutes before start.
  • Outcomes: updated labels/severities, explicit owner per P0/P1, "needs-info" pings sent.

Suggested agenda (≤15 minutes):

  1. Scan new items → label + severity.
  2. Review open P0/P1 → confirm owners and ETAs.
  3. Sweep needs‑info → send/refresh pings; close stale if justified.

Dashboards

  • Saved views for: P0 open, P1 open, needs-info, and by area.

Accessing views in Plan Manager

  • Plan overview: run report plan.
  • Backlog by priority: run list_tasks and scan for priority (0 is highest).
  • Needs‑info and area: reflect labels in titles/descriptions; group via stories.

Note: First‑class filters are planned in future iterations.