lindy-incident-runbook

# Lindy Incident Runbook

## Overview
Incident response procedures for Lindy AI integration issues.

## Prerequisites
- Access to Lindy dashboard
- Monitoring dashboards available
- Escalation contacts known
- Admin access to production

## Incident Severity Levels

| Severity | Description | Response Time | Examples |
|----------|-------------|---------------|----------|
| SEV1 | Complete outage | 15 minutes | All agents failing |
| SEV2 | Partial outage | 30 minutes | One critical agent down |
| SEV3 | Degraded | 2 hours | High latency, some errors |
| SEV4 | Minor | 24 hours | Cosmetic issues |

## Quick Diagnostics

### Step 1: Check Lindy Status
```bash
# Check Lindy status page
curl -s https://status.lindy.ai/api/v1/status | jq '.status'

# Check API health
curl -s -o /dev/null -w "%{http_code}" \
  -H "Authorization: Bearer $LINDY_API_KEY" \
  https://api.lindy.ai/v1/health
```

### Step 2: Verify Authentication
```bash
# Test API key
curl -s -H "Authorization: Bearer $LINDY_API_KEY" \
  https://api.lindy.ai/v1/users/me | jq '.email'
```

### Step 3: Check Rate Limits
```bash
# Check rate limit headers
curl -sI -H "Authorization: Bearer $LINDY_API_KEY" \
  https://api.lindy.ai/v1/users/me | grep -i "x-ratelimit"
```

## Common Incidents

### Incident: Complete API Outage

**Symptoms:**
- All API calls failing
- 5xx errors from Lindy

**Runbook:**
```markdown
1. [ ] Check https://status.lindy.ai
2. [ ] Verify it's not a local network issue
3. [ ] Check if other services on same network work
4. [ ] Enable fallback mode if available
5. [ ] Notify stakeholders
6. [ ] Open support ticket with Lindy
7. [ ] Monitor status page for updates
```

**Fallback Code:**
```typescript
async function runWithFallback(agentId: string, input: string) {
  try {
    return await lindy.agents.run(agentId, { input });
  } catch (error: any) {
    if (error.status >= 500) {
      // Enable fallback mode
      return {
        output: 'Service temporarily unavailable. Please try again