Imported from jeremylongshore/tons-of-skills-marketplace (plugins/saas-packs/firecrawl-pack/skills/firecrawl-incident-runbook/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill firecrawl-incident-runbook. Copyright stays with the author (MIT).

Firecrawl Incident Runbook

Overview

Rapid incident response procedures for Firecrawl integration failures. Covers API outage triage, credential issues, credit exhaustion, crawl job failures, and webhook delivery problems.

Prerequisites

• A declared incident commander, communications owner, escalation path, and approved secure evidence location.
• Current service ownership, target-policy, credential-revocation, and rollback references.
• A way to collect aggregate health evidence without exposing captured pages or keys.

Instructions

1. Assign severity and commander, record the start time and opaque incident ID, and stabilize unsafe jobs by disabling or rate-limiting them.
2. Determine whether the event concerns availability, credentials, policy enforcement, budget exhaustion, or data exposure; preserve redacted evidence only.
3. Apply the smallest safe mitigation, validate both recovery and a safe failure path, and communicate the user impact through the incident channel.
4. Rotate or revoke credentials when exposure is possible, and do not resume queues until policy and idempotency checks pass.
5. Record root cause, corrective actions, owners, and a follow-up review date before resolving the incident.

Output

Produce an incident receipt with severity, opaque ID, timeline, impact, mitigation, recovery verification, rollback/revocation decisions, owner, and follow-ups. Sensitive evidence remains in the approved incident store.

Examples

During a synthetic provider outage, pause the worker, verify queued jobs do not replay side effects, and send a redacted status update. Restore one canary after health recovers, then reopen normal processing only after the commander records the recovery receipt.

Severity Levels

Quick Triage (Run First)

set -euo pipefail
# 1. Test Firecrawl API directly
echo "=== API Health ==="
curl -s -w "\nHTTP %{http_code}\n" https://api.firecrawl.dev/v1/scrape \
  -H "Authorization: Bearer $FIRECRAWL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","formats":["markdown"]}' | jq '{success, error}'

# 2. Check credit balance
echo "=== Credits ==="
curl -s https://api.firecrawl.dev/v1/team/credits \
  -H "Authorization: Bearer $FIRECRAWL_API_KEY" | jq .

# 3. Check our app health
echo "=== App Health ==="
curl -sf https://api.yourapp.com/health | jq '.services.firecrawl' || echo "App unhealthy"

Decision Tree

Firecrawl API returning errors?
├─ 401: API key invalid
│   → Verify key at firecrawl.dev/app, rotate if needed
├─ 402: Credits exhausted
│   → Upgrade plan or wait for monthly reset
├─ 429: Rate limited
│   → Reduce concurrency, enable backoff, check Retry-After
├─ 500/503: Firecrawl outage
│   → Enable fallback mode, monitor firecrawl.dev status
└─ API working fine
    └─ Our integration issue
        ├─ Empty markdown → Increase waitFor, check target site
        ├─ Crawl stuck → Check job status, enforce timeout
        └─ Webhook not firing → Verify endpoint, check signature

Immediate Actions by Error Type

401 — Authentication Failure

set -euo pipefail
# Verify current key
echo "Key prefix: ${FIRECRAWL_API_KEY:0:5}"
echo "Key length: ${#FIRECRAWL_API_KEY}"

# Test with explicit key
curl -s https://api.firecrawl.dev/v1/scrape \
  -H "Authorization: Bearer $FIRECRAWL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","formats":["markdown"]}' | jq .success

# If fails: regenerate key at firecrawl.dev/app and update all environments

402 — Credits Exhausted

set -euo pipefail
# Check balance
curl -s https://api.firecrawl.dev/v1/team/credits \
  -H "Authorization: Bearer $FIRECRAWL_API_KEY" | jq .

# Immediate: disable non-critical scraping
# Long-term: upgrade plan or implement credit budget

429 — Rate Limited

// Enable emergency rate limiting
const EMERGENCY_DELAY_MS = 5000; // 5s between requests

async function emergencyScrape(url: string) {
  await new Promise(r => setTimeout(r, EMERGENCY_DELAY_MS));
  return firecrawl.scrapeUrl(url, { formats: ["markdown"] });
}

500/503 — Firecrawl Outage

// Enable graceful degradation
async function scrapeWithFallback(url: string) {
  try {
    return await firecrawl.scrapeUrl(url, { formats: ["markdown"] });
  } catch (error: any) {
    if (error.statusCode >= 500) {
      console.error("Firecrawl unavailable — using cached content");
      return getCachedContent(url); // serve stale data
    }
    throw error;
  }
}

Communication Templates

Internal (Slack)

P[1-4] INCIDENT: Firecrawl Integration
Status: INVESTIGATING
Impact: [Describe user-facing impact]
Error: [401/402/429/500] — [brief description]
Action: [What you're doing right now]
Next update: [time]

Post-Incident

Evidence Collection

set -euo pipefail
# Collect debug bundle
mkdir -p incident-$(date +%Y%m%d)
curl -s https://api.firecrawl.dev/v1/team/credits \
  -H "Authorization: Bearer $FIRECRAWL_API_KEY" > incident-$(date +%Y%m%d)/credits.json

# Application logs
kubectl logs -l app=my-app --since=1h | grep -i firecrawl > incident-$(date +%Y%m%d)/logs.txt 2>/dev/null || true

Postmortem Template

## Incident: Firecrawl [Error Type]
Date: YYYY-MM-DD | Duration: X hours | Severity: P[1-4]

### Summary
[1-2 sentence description]

### Timeline
- HH:MM — [First alert]
- HH:MM — [Investigation started]
- HH:MM — [Root cause identified]
- HH:MM — [Resolved]

### Root Cause
[Technical explanation]

### Action Items
- [ ] [Preventive measure] — Owner — Due date

Error Handling

Resources

• Firecrawl Dashboard
• Firecrawl Status
• GitHub Issues

Next Steps

For data handling, see firecrawl-data-handling.