coderabbit-observability

# CodeRabbit Observability

## Overview
Monitor CodeRabbit AI code review effectiveness, review latency, and team adoption. Key metrics include time-to-first-review (how fast CodeRabbit posts after PR creation), comment acceptance rate (comments resolved vs dismissed), review coverage (percentage of PRs reviewed), and per-repository review volume.

## Prerequisites
- CodeRabbit installed on GitHub/GitLab organization
- GitHub CLI (`gh`) authenticated with org access
- Access to CodeRabbit dashboard at app.coderabbit.ai

## Key Metrics

| Metric | Target | Why It Matters |
|--------|--------|----------------|
| Review coverage | > 90% | PRs without review = blind spots |
| Time-to-review | < 5 min | Fast feedback keeps developers in flow |
| Comment acceptance | > 40% | Low acceptance = noisy reviews |
| Comments per PR | 3-8 | Too many = fatigue, too few = not useful |
| Review state: APPROVED | > 60% | High approval = clean code culture |

## Instructions

### Step 1: Measure Review Coverage
```bash
#!/bin/bash
# coderabbit-coverage.sh - Review coverage for a repo
set -euo pipefail

ORG="${1:?Usage: $0 <org> <repo> [days]}"
REPO="${2:?Usage: $0 <org> <repo> [days]}"
DAYS="${3:-30}"

echo "=== CodeRabbit Review Coverage ==="
echo "Repository: $ORG/$REPO"
echo "Period: Last $DAYS days"
echo ""

TOTAL=0
REVIEWED=0
APPROVED=0
CHANGES_REQUESTED=0

SINCE=$(date -d "$DAYS days ago" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null || date -v-${DAYS}d +%Y-%m-%dT%H:%M:%SZ)

for PR_NUM in $(gh api "repos/$ORG/$REPO/pulls?state=all&per_page=50&sort=created&direction=desc" \
  --jq ".[] | select(.created_at > \"$SINCE\") | .number"); do

  TOTAL=$((TOTAL + 1))

  CR_STATE=$(gh api "repos/$ORG/$REPO/pulls/$PR_NUM/reviews" \
    --jq '[.[] | select(.user.login=="coderabbitai[bot]")] | last | .state // "none"' 2>/dev/null || echo "none")

  if [ "$CR_STATE" != "none" ] && [ "$CR_STATE" != "null" ]; then
    REVIEWED=$((REVIEWED + 1))
    [ "$CR_STATE" = "APPROVED" ] && APPROVED=$((APPROVED + 1))
    [ "$CR_STATE" = "CHANGES_REQUESTED" ] && CHANGES_REQUESTED=$((CHANGES_REQUESTED + 1))
  fi
done

if [ "$TOTAL" -gt 0 ]; then
  echo "Total PRs: $TOTAL"
  echo "Reviewed by CodeRabbit: $REVIEWED ($(( REVIEWED * 100 / TOTAL ))%)"
  echo "  Approved: $APPROVED"
  echo "  Changes Requested: $CHANGES_REQUESTED"
else
  echo "No PRs found in the last $DAYS days"
fi
```

### Step 2: Track Comment Volume and Acceptance
```bash
set -euo pipefail
ORG="${1:-your-org}"
REPO="${2:-your-repo}"

echo "=== CodeRabbit Comment Analysis ==="
echo ""

TOTAL_COMMENTS=0
PR_COUNT=0

for PR_NUM in $(gh api "repos/$ORG/$REPO/pulls?state=closed&per_page=20" --jq '.[].number'); do
  COMMENTS=$(gh api "repos/$ORG/$REPO/pulls/$PR_NUM/comments" \
    --jq '[.[] | select(.user.login=="coderabbitai[bot]")] | length' 2>/dev/null || echo "0")

  if [ "$COMMENTS" -gt 0 ]; then
    TOTAL_COMMENTS=$((TOTAL_COMMENTS + COMMENTS))
    PR_COUNT=$((PR_COUNT + 1))
    echo "PR #$PR_NUM: $COMMENTS comments"
  fi
done

if [ "$PR_COUNT" -gt 0 ]; then
  echo ""
  echo "Average comments per PR: $(( TOTAL_COMMENTS / PR_COUNT ))"
  echo ""
  echo "Healthy ranges:"
  echo "  1-3 comments/PR → Profile may be too chill"
  echo "  3-8 comments/PR → Good signal-to-noise ratio"
  echo "  10+ comments/PR → Consider switching to chill profile"
fi
```

### Step 3: Build a GitHub Actions Dashboard
```yaml
# .github/workflows/coderabbit-metrics.yml
name: CodeRabbit Weekly Metrics

on:
  schedule:
    - cron: '0 9 * * 1'    # Every Monday at 9 AM UTC
  workflow_dispatch:         # Manual trigger

jobs:
  metrics:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/github-script@v7
        with:
          script: |
            const { data: pulls } = await github.rest.pulls.list({
              owner: context.repo.owner,
              repo: context.repo.repo,
              state: 'closed',
              per_page: 50,
              sort: 'updated',
              direction: 'desc',
            });

            let reviewed = 0;
            let approved = 0;
            let changesRequested = 0;
            let totalComments = 0;

            for (const pr of pulls) {
              const { data: reviews } = await github.rest.pulls.listReviews({
                owner: context.repo.owner,
                repo: context.repo.repo,
                pull_number: pr.number,
              });

              const crReview = reviews.find(r => r.user.login === 'coderabbitai[bot]');
              if (crReview) {
                reviewed++;
                if (crReview.state === 'APPROVED') approved++;
                if (crReview.state === 'CHANGES_REQUESTED') changesRequested++;
              }

              const { data: comments } = await github.rest.pulls.listReviewComments({
                owner: context.repo.owner,
                repo: context.repo.repo,
                pull_number: pr.number,
              });
              totalComments += comments.filter(c => c.user.login === 'coderabbitai[bot]').length;
            }

            const summary = [
              `## CodeRabbit Weekly Metrics`,
              `- **Coverage**: ${reviewed}/${pulls.length} PRs reviewed (${Math.round(reviewed/pulls.length*100)}%)`,
              `- **Approved**: ${approved}`,
              `- **Changes Requested**: ${changesRequested}`,
              `- **Avg Comments/PR**: ${reviewed > 0 ? Math.round(totalComments/reviewed) : 0}`,
            ].join('\n');

            core.summary.addRaw(summary).write();
            core.info(summary);
```

### Step 4: Set Up Alerts for Review Gaps
```yaml
# .github/workflows/coderabbit-alert.yml
name: CodeRabbit Review Alert

on:
  pull_request:
    types: [opened]

jobs:
  check-review-expected:
    runs-on: ubuntu-latest
    steps:
      - name: Wait for CodeRabbit review
        uses: actions/github-script@v7
        with:
          script: |
            // Wait 10 minutes, then check if CodeRabbit reviewed
            await new Promise(r => setTimeout(r, 600000));

            const { data: reviews } = await github.rest.pulls.listReviews({
              owner: context.repo.owner,
              repo: context.repo.repo,
              pull_number: context.issue.number,
            });

            const crReview = reviews.find(r => r.user.login === 'coderabbitai[bot]');

            if (!crReview) {
              core.warning(
                'CodeRabbit has not reviewed this PR after 10 minutes. ' +
                'Check: App installation, .coderabbit.yaml, base_branches config.'
              );
            }
```

### Step 5: CodeRabbit Dashboard Summary
```markdown
# Build a summary dashboard with these data points:

## Weekly Dashboard Template

| Metric | This Week | Last Week | Trend |
|--------|-----------|-----------|-------|
| PRs opened | | | |
| PRs reviewed by CR | | | |
| Coverage % | | | |
| Avg comments/PR | | | |
| Approval rate | | | |
| Time to first review | | | |

## Action Items:
- Coverage < 90%: Check App installation, base_branches config
- Avg comments > 10: Switch to "chill" profile
- Avg comments < 2: Switch to "assertive" profile
- Approval rate < 50%: Review path_instructions for relevance
```

## Output
- Review coverage metrics calculated per repository
- Comment volume and acceptance rate tracked
- Weekly metrics GitHub Action workflow
- Alert workflow for missing reviews
- Dashboard template for team reporting

## Error Handling
| Issue | Cause | Solution |
|-------|-------|----------|
| Coverage below 90% | Some PRs not reviewed | Check `base_branches` and `ignore_title_keywords` |
| Low acceptance rate | Too many false positives | Tune `path_instructions` and switch to `chill` |
| No metrics data | No closed PRs in period | Extend the time window |
| API rate limited | Too many `gh api` calls | Add pagination and caching |

## Resources
- [CodeRabbit Dashboard](https://app.coderabbit.ai)
- [GitHub REST API - Pulls](https://docs.github.com/en/rest/pulls)
- [GitHub Actions Job Summaries](https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions#adding-a-job-summary)

## Next Steps
For incident response, see `coderabbit-incident-runbook`.