# 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

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

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

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

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 later.',
        fallback: true,
      };
    }
    throw error;
  }
}

1. [ ] Check current usage in dashboard
2. [ ] Identify spike source (which agent/automation)
3. [ ] Reduce request rate or implement throttling
4. [ ] Consider upgrading plan if legitimate traffic
5. [ ] Implement request queuing

const queue = new PQueue({ concurrency: 5, interval: 1000, intervalCap: 10 });

async function throttledRun(agentId: string, input: string) {
  return queue.add(() => lindy.agents.run(agentId, { input }));
}

1. [ ] Identify affected agent(s)
2. [ ] Check agent configuration hasn't changed
3. [ ] Review recent runs for patterns
4. [ ] Test with simple input
5. [ ] Check if tools are working
6. [ ] Rollback to previous version if needed

async function diagnoseAgent(agentId: string) {
  const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });

  // Get agent details
  const agent = await lindy.agents.get(agentId);
  console.log('Agent:', agent.name, agent.status);

  // Check recent runs
  const runs = await lindy.runs.list({ agentId, limit: 10 });
  const failures = runs.filter((r: any) => r.status === 'failed');
  console.log(`Failures: ${failures.length}/${runs.length}`);

  // Test run
  try {
    const test = await lindy.agents.run(agentId, { input: 'Hello' });
    console.log('Test run: SUCCESS');
  } catch (e: any) {
    console.log('Test run: FAILED -', e.message);
  }

  return { agent, runs, failures };
}

1. [ ] Check Lindy status page for degradation
2. [ ] Review latency metrics by agent
3. [ ] Check if issue is with specific agent
4. [ ] Verify instructions aren't causing long responses
5. [ ] Consider reducing max_tokens
6. [ ] Implement streaming if not already

## Incident Report: [Title]

**Date:** YYYY-MM-DD
**Duration:** X hours Y minutes
**Severity:** SEV1/2/3/4
**Impact:** [Description of user impact]

### Timeline
- HH:MM - Incident detected
- HH:MM - On-call paged
- HH:MM - Root cause identified
- HH:MM - Resolution applied
- HH:MM - All clear

### Root Cause
[What caused the incident]

### Resolution
[What fixed it]

### Action Items
- [ ] [Preventive action 1]
- [ ] [Preventive action 2]