Metrics, Monitoring & Logging (New Relic & Sentry)

Overview

Learnille utilizes New Relic as its primary Observability, Application Performance Monitoring (APM), and Log Management platform, combined with Sentry for real-time exception tracking.

📊 1. Observability Stack Architecture

                                [ Learnille Services ]
                                           |
         +---------------------------------+---------------------------------+
         |                                 |                                 |
         v                                 v                                 v
[ NestJS Backend API ]          [ Vite React Frontends ]          [ Background Workers / DB ]
  - New Relic Node.js Agent       - Sentry Browser SDK              - New Relic Infrastructure
  - Winston New Relic Transport   - Web Vitals Tracking             - Redis & Postgres Metrics
         |                                 |                                 |
         +---------------------------------+---------------------------------+
                                           |
                                           v
                       +---------------------------------------+
                       |          New Relic One Hub            |
                       |  - APM & Transaction Tracing          |
                       |  - Centralized Log Analytics          |
                       |  - Real-time Alerting & Dashboards    |
                       +---------------------------------------+

🛠 2. New Relic Integration Setup

Node.js Backend APM Configuration

In the NestJS server (server/src/main.ts and newrelic.js):

exports.config = {
  app_name: ['Learnille API (Self-Hosted)'],
  license_key: process.env.NEW_RELIC_LICENSE_KEY,
  logging: {
    level: 'info'
  },
  allow_all_headers: true,
  attributes: {
    exclude: [
      'request.headers.cookie',
      'request.headers.authorization'
    ]
  }
};

Structured Log Transport

Winston logger routes all server logs directly to New Relic using the official @newrelic/winston-enricher transport.

🚨 3. Alerting Thresholds & Dashboards

Metric	Target SLA	New Relic Alert Condition
API Error Rate	< 0.5%	Alert triggered if HTTP 5xx rate > 1% for 5 mins.
P95 Transaction Duration	< 250ms	Alert triggered if average response time > 500ms for 5 mins.
Database Query Latency	< 20ms	Alert triggered if SQL query duration > 100ms.
Host CPU / Memory Utilization	< 75%	Alert triggered if host RAM usage > 85%.