Oncall review

On-Call Quality Review

Evaluate the quality of a completed on-call (or first-responder) shift by scoring against a rubric. Produces a star rating and qualitative assessment.

When to use

• "How was this week's on-call?", "rate the FR shift", "on-call quality check"
• As part of periodic team health reviews

Prerequisites

• A GitHub issue tracking the on-call shift (typically auto-created with a checklist template)
• Access to the issue via gh CLI
• A scoring rubric stored in the vault (location from setup config)

Process

Step 1: Fetch the shift issue

gh issue view {number} --repo {org}/{team-repo} --json title,body,comments,labels,assignees,createdAt,closedAt

Extract:

• Assignee: who was on-call
• Checklist items: count total, count checked
• Comments: count total, assess quality (see below)
• Handoff: look for a handoff comment or section

Step 2: Score against rubric

Apply the scoring rubric. A reasonable default rubric:

Criterion	Points	How to assess
Checklist completion ≥ 20%	+1	Count checked / total checkboxes
Checklist completion ≥ 50%	+1	Same
Comments ≥ 3	+1	Count non-bot comments
Comments ≥ 6	+1	Same
Handoff completed	+1	Look for a handoff comment, section, or linked handoff issue

Maximum score: ⭐⭐⭐⭐⭐

Teams should customize this rubric to match their expectations. Store it in the vault (e.g., Areas/Team Metrics/On-Call Quality Rubric.md).

Step 3: Qualitative assessment

Beyond the raw score, assess:

• Investigation quality: Did comments show actual debugging (Splunk queries, Sentry links, SLO analysis)? Or just "acknowledged, auto-resolved"?
• Checklist engagement: Were unchecked items skipped because they didn't apply, or because they were missed?
• Handoff quality: Does the handoff give the next person enough context to start without spelunking? Or is it "nothing happened this week"?
• Incident response: If incidents fired, were they triaged promptly? Is there follow-up documented?

Step 4: Output

Write a summary (to the vault or inline, depending on context):

### On-Call Review: @{assignee} ({date_range})

**Score: ⭐⭐⭐** (3/5)

| Criterion | Result |
|-----------|--------|
| Checkboxes ≥ 20% | ✅ (47%) |
| Checkboxes ≥ 50% | ❌ (47%) |
| Comments ≥ 3 | ✅ (3) |
| Comments ≥ 6 | ❌ (3) |
| Handoff | ⚠️ Partial |

**Notes:**
- Investigation quality was above average (detailed SLO breakdown with monitoring links)
- Checkbox completion hurt by [reason if identifiable]
- Handoff was [complete/partial/missing]: [specifics]

If maintaining a tracking doc, append this to the historical record.

Rubric customization

The default rubric above is a starting point. Teams should adjust based on:

• Shift type: On-call vs. first-responder may have different expectations
• Team size: Smaller teams may have fewer incidents and lower comment baselines
• Maturity: New teams should start with lower thresholds and raise them over time

Store the active rubric in the vault so the team can see and iterate on it.

Rules

• Score the shift, not the person. The review assesses the quality of the work product, not the individual. Frame feedback constructively.
• Account for quiet weeks. A low score during a genuinely quiet week (no incidents, no alerts) should be noted as context, not penalized.
• Flag data issues. If the issue body appears truncated (common with long checklists), note which criteria couldn't be assessed.
• Be honest. A 2-star shift is a 2-star shift. Don't inflate scores to be nice, but do explain what would have made it better.