Monitoring & Observability

AdRecon’s production monitoring spans Vercel serverless functions, Supabase database metrics, and custom rate limiting logs.

Monitoring Stack

Vercel Logs

Function execution logs, deployment history, and edge network metrics

Supabase Metrics

Query performance, connection pools, RLS policy hits, and storage usage

Admin Dashboard

Real-time user stats, saved ads, projects, and feed health

Rate Limit Logs

Page Ripper usage tracking via page_rip_log table

Vercel Monitoring

Function Logs

Access real-time serverless function logs in the Vercel dashboard:

Navigate to Project Logs

Go to Vercel Dashboard → Your Project → Logs

Filter by Function

Use the function filter to isolate specific endpoints:

/api/download-page — Page Ripper captures
/api/admin/users — Admin user management
/api/ad-media — Media proxy requests
/api/admin/scrape — Admin scraping operations
/api/landing-ripper/webhook — Webhook events

Search Error Patterns

Filter logs by severity:

Error: Function threw an exception
Warning: Non-critical issues (rate limits, timeouts)
Info: Standard execution logs

Key Metrics

Vercel provides these critical metrics per function:

Invocation count: Total requests per time window
Error rate: Percentage of failed invocations
Duration (p50, p95, p99): Response time percentiles
Memory usage: Peak RAM consumption
Timeout rate: Functions exceeding max duration

The /api/download-page function has a 120-second timeout (configured in vercel.json:8). Monitor timeout rates — values above 5% indicate sites with heavy lazy-loading or slow networks.

Function Timeout Configuration

These functions have extended timeouts due to expensive operations:

vercel.json

{
  "functions": {
    "api/download-page.js": {
      "maxDuration": 120
    },
    "api/admin/scrape.js": {
      "maxDuration": 60
    },
    "api/landing-ripper/jobs.js": {
      "maxDuration": 60
    },
    "api/landing-ripper/webhook.js": {
      "maxDuration": 120
    },
    "api/landing-ripper/download/[token].js": {
      "maxDuration": 60
    }
  }
}

Monitor these functions closely — hitting the max duration triggers a 504 Gateway Timeout and burns execution time budget.

Supabase Monitoring

Database Metrics

Access Supabase metrics in the project dashboard:

Open Database Reports

Supabase Dashboard → Your Project → Database → Reports

Monitor Query Performance

Track slow queries:

Top slow queries: Queries exceeding 100ms
Most frequent queries: High-traffic feed lookups
Index usage: Verify indexes are hit for ads_feed_v2 scans

Check Connection Pool

Monitor active connections:

Active connections: Current open sessions
Idle connections: Pooled but unused
Max connections: Hard limit (default: 60 for free tier)

Critical Tables to Monitor

-- Monitor row count growth
SELECT COUNT(*) FROM ads_feed_v2;

-- Check most recent ad ingestion timestamp
SELECT MAX(ad_delivery_start_time) FROM ads_feed_v2;

RLS Policy Performance

Row-Level Security (RLS) policies can impact query performance. Monitor policy hit rates:

-- Check if queries are scanning full tables despite RLS
EXPLAIN ANALYZE
SELECT * FROM user_saved_ads
WHERE user_id = 'some-user-id';

Look for:

Index Scan (good) vs Seq Scan (bad)
Rows Removed by Filter — high values indicate missing indexes

If user_saved_ads queries show sequential scans, verify the user_id column has an index.

Admin Dashboard Stats

The admin dashboard (/admin or /app/admin) provides real-time operational metrics.

Available Metrics

Accessed via GET /api/admin/users, the admin API returns:

Admin Stats Response

{
  "stats": {
    "totalUsers": 142,
    "adminUsers": 3,
    "recentlyActiveUsers7d": 38,
    "confirmedUsers": 128,
    "savedAds": 1847,
    "projects": 93,
    "indexedAds": 45203
  }
}

These stats are cached for 60 seconds (see api/admin/users.js:21) to reduce database load during repeated admin dashboard views.

Metric Definitions

totalUsers: All registered accounts (confirmed + unconfirmed)
adminUsers: Users with app_metadata.user_type = 'admin'
recentlyActiveUsers7d: Users with last_sign_in_at within 7 days
confirmedUsers: Users with email_confirmed_at set
savedAds: Total rows in user_saved_ads table
projects: Total rows in user_projects table
indexedAds: Total rows in ads_feed_v2 view

Cache Invalidation

The admin API caches two datasets:

// Cached for 15 seconds
const USER_CACHE_TTL_MS = 15_000;

// Invalidated on user mutations (PATCH, DELETE)
invalidateUserCache();

Cache invalidation happens automatically when:

Admin promotes/demotes a user (PATCH /api/admin/users)
Admin deletes a user (DELETE /api/admin/users)

Page Ripper Rate Limiting

The Page Ripper feature includes per-user rate limiting to prevent abuse.

Rate Limit Configuration

Defined in api/download-page.js:52-53:

const RATE_LIMIT_MAX = 10;
const RATE_LIMIT_WINDOW_MINUTES = 15;

This allows 10 page captures per user per 15-minute window.

Monitoring Rate Limits

Query recent activity:

page_rip_log Analysis

-- Users hitting rate limit in last 15 minutes
SELECT user_id, COUNT(*) as rip_count
FROM page_rip_log
WHERE created_at >= NOW() - INTERVAL '15 minutes'
GROUP BY user_id
HAVING COUNT(*) >= 10
ORDER BY rip_count DESC;

The rate limiter fails open — if the page_rip_log table is missing or queries error, requests proceed without rate limiting (see api/download-page.js:249).

Rate Limit Response

When a user exceeds the limit, the API returns:

429 Too Many Requests

{
  "error": "Rate limit exceeded. Maximum 10 page captures per 15 minutes."
}

HTTP headers:

Retry-After: 900 (15 minutes in seconds)

Historical Analysis

Track Page Ripper usage trends:

Historical Rip Activity

-- Daily rip volume (last 30 days)
SELECT
  DATE(created_at) as date,
  COUNT(*) as total_rips,
  COUNT(DISTINCT user_id) as unique_users
FROM page_rip_log
WHERE created_at >= NOW() - INTERVAL '30 days'
GROUP BY DATE(created_at)
ORDER BY date DESC;

Error Tracking Patterns

Common Error Signatures

// Logged when user attempts to rip internal address
console.error('[download-page] SSRF attempt blocked:', hostname);

// Response: 400 Bad Request
// "Requests to private/internal addresses are not allowed."

Logging Best Practices

All serverless functions follow this pattern:

try {
  // ... operation ...
} catch (error) {
  const message = error instanceof Error
    ? error.message
    : 'Unknown error';
  console.error('[function-name] Context:', message);
  json(res, 500, { error: message });
}

Error messages include:

Context prefix: [function-name] for easy log filtering
Stack traces: Automatically captured by Vercel
User-facing message: Sanitized error returned in JSON response

Alerts and Thresholds

Recommended monitoring alerts:

Vercel Function Timeouts

Alert when /api/download-page timeout rate exceeds 10% over 1 hour.

Supabase Connection Pool

Alert when active connections exceed 80% of max connections.

Feed Staleness

Alert when MAX(ad_delivery_start_time) in ads_feed_v2 is older than 24 hours.

Admin API Error Rate

Alert when /api/admin/users error rate exceeds 5% over 15 minutes.

Vercel Pro and Enterprise plans support custom alerting via integrations with Datadog, New Relic, and other APM tools.

Performance Baselines

Normal operating ranges:

/api/ad-media (p95): < 200ms
/api/admin/users (p95): < 500ms (uncached), < 50ms (cached)
/api/download-page (p50): 15-30 seconds
ads_feed_v2 query (p95): < 300ms
user_saved_ads insert: < 50ms

Baseline times assume healthy Supabase instance. If p95 query times exceed 1 second, investigate missing indexes or RLS policy overhead.

Setup

Operations

Monitoring & Observability

Monitoring Stack

Vercel Logs

Supabase Metrics

Admin Dashboard

Rate Limit Logs

Vercel Monitoring

Function Logs

Key Metrics

Function Timeout Configuration

Supabase Monitoring

Database Metrics

Critical Tables to Monitor

RLS Policy Performance

Admin Dashboard Stats

Available Metrics

Metric Definitions

Cache Invalidation

Page Ripper Rate Limiting

Rate Limit Configuration

Monitoring Rate Limits

Rate Limit Response

Historical Analysis

Error Tracking Patterns

Common Error Signatures

Logging Best Practices

Alerts and Thresholds

Performance Baselines

Build docs developers (and LLMs) love

Setup

Operations

​Monitoring Stack

Vercel Logs

Supabase Metrics

Admin Dashboard

Rate Limit Logs

​Vercel Monitoring

​Function Logs

​Key Metrics

​Function Timeout Configuration

​Supabase Monitoring

​Database Metrics

​Critical Tables to Monitor

​RLS Policy Performance

​Admin Dashboard Stats

​Available Metrics

​Metric Definitions

​Cache Invalidation

​Page Ripper Rate Limiting

​Rate Limit Configuration

​Monitoring Rate Limits

​Rate Limit Response

​Historical Analysis

​Error Tracking Patterns

​Common Error Signatures

​Logging Best Practices

​Alerts and Thresholds

​Performance Baselines

Build docs developers (and LLMs) love

Monitoring Stack

Vercel Monitoring

Function Logs

Key Metrics

Function Timeout Configuration

Supabase Monitoring

Database Metrics

Critical Tables to Monitor

RLS Policy Performance

Admin Dashboard Stats

Available Metrics

Metric Definitions

Cache Invalidation

Page Ripper Rate Limiting

Rate Limit Configuration

Monitoring Rate Limits

Rate Limit Response

Historical Analysis

Error Tracking Patterns

Common Error Signatures

Logging Best Practices

Alerts and Thresholds

Performance Baselines