canva-incident-runbook

# Canva Incident Runbook ## Overview Rapid incident response for Canva Connect API integration failures. Covers triage, mitigation, escalation, and postmortem. ## Quick Triage (First 5 Minutes) ```bash #!/bin/bash # canva-triage.sh — Run immediately when incident detected echo "=== Canva Triage ===" # 1. Is it Canva or us? echo -n "Canva API: " curl -s -o /dev/null -w "HTTP %{http_code} (%{time_total}s)\n" \ -H "Authorization: Bearer $CANVA_ACCESS_TOKEN" \ "https://api.canva.com/rest/v1/users/me" # 2. Check our health endpoint echo -n "Our health: " curl -s -o /dev/null -w "HTTP %{http_code}\n" \ "https://api.ourapp.com/health" # 3. Error rate (if Prometheus available) echo "Error rate (5min):" curl -s "localhost:9090/api/v1/query?query=rate(canva_api_errors_total[5m])" \ | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['data']['result'])" 2>/dev/null \ || echo "Prometheus not available" # 4. Rate limit status echo -n "Rate limit remaining: " curl -sD - -o /dev/null -H "Authorization: Bearer $CANVA_ACCESS_TOKEN" \ "https://api.canva.com/rest/v1/designs?limit=1" 2>&1 \ | grep -i "x-ratelimit-remaining" || echo "unknown" ``` ## Decision Tree ``` API returning errors? ├── YES → What HTTP status? │ ├── 401 → Token expired → Refresh token, check rotation │ ├── 403 → Scope issue → Verify integration permissions │ ├── 429 → Rate limited → Enable backoff, check Retry-After │ ├── 5xx → Canva outage → Enable fallback, monitor status page │...