Información y análisis de logs

Supervisión y resolución de problemas en AWS

John Q. Martin

Principal Consultant

Estructura de consultas de Logs Insights

Estructura de consultas de Logs Insights que muestra los comandos encadenados con pipes: fields, filter, stats, sort y limit

Lenguajes de consulta: misma consulta, tres formas

CloudWatch Logs Insights QL

fields @timestamp, @message, level
| filter level = "ERROR"
| filter @timestamp > ago(1h)
| sort @timestamp desc
| limit 20

OpenSearch PPL

source = `/aws/lambda/my-function`
| where level = 'ERROR' and @timestamp > ago(1h)
| fields @timestamp, @message, level
| sort - @timestamp
| head 20

Lenguajes de consulta: OpenSearch SQL

OpenSearch SQL

SELECT `@timestamp`, `@message`, level
FROM `/aws/lambda/my-function`
WHERE level = 'ERROR'
  AND `@timestamp` > ago(1h)
ORDER BY `@timestamp` DESC
LIMIT 20

Sintaxis estándar SELECT / FROM / WHERE
Añade JOIN entre grupos de logs
Elige lo que mejor encaje con tu equipo

Escenario 1: Picos de errores

fields @timestamp, @message, level
| filter level = "ERROR"
| stats count() as error_count by bin(5m)
| sort @timestamp desc

Mejorado: agrupar por tipo de error

fields @timestamp, error_type, error_message
| filter level = "ERROR"
| stats count() as count by error_type, bin(5m)
| sort count desc

Escenario 2: Endpoints de API lentos

fields @timestamp, endpoint, response_time, status_code
| filter response_time > 1000
| stats avg(response_time) as avg_time,
        max(response_time) as max_time,
        count() as slow_requests
        by endpoint
| sort avg_time desc

Análisis en detalle

fields @timestamp, endpoint, response_time, user_id, request_id
| filter endpoint = "/api/users" and response_time > 1000
| sort response_time desc
| limit 20

Escenario 3: Autenticación fallida

fields @timestamp, user_id, ip_address, action
| filter action = "login_failed"
| stats count() as failed_attempts by user_id, ip_address
| sort failed_attempts desc
| limit 50

Análisis temporal

fields @timestamp, user_id, ip_address
| filter action = "login_failed"
| stats count() as attempts by ip_address, bin(1h)
| filter attempts > 10
| sort attempts desc

Escenarios 4 y 5: timeouts de BD y fugas de memoria

Timeouts de base de datos

fields @timestamp, @message
| filter @message like /database/ and @message like /timeout|error|failed/
| parse @message "timeout after * seconds" as timeout_duration
| stats count() as timeout_count,
        avg(timeout_duration) as avg_timeout
        by bin(5m)

Detección de fugas de memoria

fields @timestamp, memory_used_mb, heap_size_mb
| stats avg(memory_used_mb) as avg_memory,
        max(memory_used_mb) as max_memory
        by bin(1h)
| sort @timestamp asc

Escenario 6: Trazado de solicitudes

fields @timestamp, @message, request_id, service, action
| filter request_id = "abc123def456"
| sort @timestamp asc

Trazado multi-servicio

fields @timestamp, service, action, duration_ms, status
| filter request_id = "abc123def456"
| sort @timestamp asc
| display @timestamp, service, action, duration_ms, status

Técnicas avanzadas: parse y Regex

Análisis de logs no estructurados

fields @timestamp, @message
| parse @message "[*] User * failed to access resource * from IP *"
    as level, user, resource, ip

Análisis con Regex

| parse @message /Request completed in (?<duration>\d+)ms with status (?<status>\d+)/

Técnicas avanzadas: campos calculados

fields @timestamp, requests, errors
| fields error_rate = (errors / requests) * 100

fields @timestamp, status_code
| fields status_category =
    case(status_code < 300, "success",
         status_code < 500, "client_error",
         status_code >= 500, "server_error")
| stats count() as request_count by status_category

Optimización de consultas

Cuatro reglas de optimización de consultas: aplicar un filtro de rango temporal antes de agregar, limitar resultados y preferir agregaciones

Detección de anomalías y correlación

Detección estadística de anomalías

Calcular línea base: media y desviación estándar en 7 días
Comparar el periodo actual con la línea base
Marcar valores > 2 desviaciones estándar como anómalos

Correlación de tasa de error

fields @timestamp, level
| stats count() as total_requests,
        sum(case(level = "ERROR", 1, 0)) as errors
        by bin(5m)
| fields error_rate = (errors / total_requests) * 100

Resumen

CloudWatch Logs Insights: servicio de consultas interactivo, sin infraestructura
Tres lenguajes: Logs Insights QL, OpenSearch PPL, OpenSearch SQL
Seis escenarios de diagnóstico: picos de error, endpoints lentos, auth fallida, timeouts de BD, fugas de memoria, trazado de solicitudes
Técnicas avanzadas: parse para texto no estructurado, bin() para series temporales, pct() para percentiles
Correlación y anomalías: análisis de tasa de error, comparación con línea base estadística
De la investigación a la automatización: filtros de métricas y alarmas cierran el ciclo

¡Vamos a practicar!

Supervisión y resolución de problemas en AWS