Monitoring & Observability

Tarasol uses a comprehensive monitoring stack for application performance, logging, and metrics collection across all microservices.

Stack Overview

Tool	Purpose
Elastic APM	Application performance monitoring
Elasticsearch	Log aggregation and search
Winston	Structured logging framework
Prometheus	Metrics collection
Morgan	HTTP request logging

Elastic APM

Every NestJS service includes the Elastic APM agent for:

Transaction Tracing — End-to-end request tracking across services
Error Tracking — Automatic error capture and alerting
Performance Metrics — Response times, throughput, latency

Configuration

Name	Type	Default	Description
ELASTIC_APM_SERVER_URL	string	-	APM Server URL
ELASTIC_APM_SERVICE_NAME	string	-	Service name for APM
ELASTIC_APM_SECRET_TOKEN	string	-	APM authentication token
ELASTIC_APM_ENVIRONMENT	string	production	Environment name

Logging

All services use Winston with ECS (Elastic Common Schema) format:

// Logging configuration
{
  format: winston.format.combine(
    winston.format.timestamp(),
    ecsFormat()
  ),
  transports: [
    new winston.transports.Console(),
    new ElasticsearchTransport({
      node: ELASTICSEARCH_URL
    })
  ]
}

Log Levels

Level	Usage
`error`	Application errors, unhandled exceptions
`warn`	Deprecation notices, retry attempts
`info`	Request lifecycle, business events
`debug`	Detailed diagnostic information

Prometheus Metrics

Services expose metrics at /metrics endpoint:

http_request_duration_seconds — Request duration histogram
http_requests_total — Total request counter
active_connections — Current active connections
rabbitmq_messages_consumed — Messages consumed counter

Health Checks

Every service exposes /health with checks for:

Database connectivity
Redis connectivity
RabbitMQ connectivity
External service dependencies