Saltar a contenido

SAFE-POINT-035 - Dashboard y alertas visibles La Directa

Fecha: 2026-06-15

Objetivo

Hacer visible en Grafana ejecutivo el monitoreo Blackbox ya creado para La Directa, sin tocar la VPS WooCommerce ni cambiar red, DNS, WordPress, WooCommerce, exporters, Prometheus publico o Alertmanager publico.

SAFE POINT inicial

  • Workspace: C:\APV\openclawai
  • Rama inicial: main
  • Estado inicial: main...origin/main
  • HEAD inicial: 3a2d4de9316f362cb873fc51c9171b72b2de2adf

Alcance ejecutado

  • Se reviso SAFE-POINT-034:
  • infra/observability/prometheus/targets/blackbox-http-targets.json
  • infra/observability/prometheus/rules/service-alerts.yml
  • infra/observability/grafana/dashboards/04-openclaw/o2-service-availability.json
  • documentacion de observabilidad y catalogos O4
  • Se creo dashboard ejecutivo especifico:
  • infra/observability/grafana/dashboards/10-executive-board/o97-ecommerce-ladirecta.json
  • titulo Grafana: Ecommerce La Directa
  • carpeta provisionada: 10 - Executive Board
  • Se mantuvo Service Availability como tablero operativo general.

Cobertura del dashboard

  • Estado https://ladirecta.com.ar/.
  • Estado https://ladirecta.ar/.
  • Uptime 24h total y por endpoint.
  • Latencia HTTP / probe_duration_seconds.
  • SSL days remaining.
  • Probe duration actual.
  • HTTP status code.
  • Alertas activas del tenant ladirecta.

Alertas La Directa validadas

  • EcommerceLaDirectaComArDown
  • EcommerceLaDirectaArDown
  • EcommerceLaDirectaHighLatency
  • EcommerceLaDirectaSslExpiryWarning
  • EcommerceLaDirectaSslExpiryCritical

Queries operativas

  • Estado ambos endpoints: probe_success{job="blackbox-http", tenant="ladirecta", service="ecommerce-ladirecta"}
  • Estado .com.ar: probe_success{job="blackbox-http", instance="https://ladirecta.com.ar/"}
  • Estado .ar: probe_success{job="blackbox-http", instance="https://ladirecta.ar/"}
  • Uptime 24h: 100 * avg_over_time(probe_success{job="blackbox-http", tenant="ladirecta", service="ecommerce-ladirecta"}[24h])
  • Latencia: probe_duration_seconds{job="blackbox-http", tenant="ladirecta", service="ecommerce-ladirecta"}
  • SSL days remaining: (probe_ssl_earliest_cert_expiry{job="blackbox-http", tenant="ladirecta", service="ecommerce-ladirecta"} - time()) / 86400
  • Alertas activas: ALERTS{tenant="ladirecta", alertstate="firing"}

Validaciones locales

  • JSON dashboard nuevo valido: infra/observability/grafana/dashboards/10-executive-board/o97-ecommerce-ladirecta.json.
  • JSON targets Blackbox valido: infra/observability/prometheus/targets/blackbox-http-targets.json.
  • YAML valido:
  • infra/observability/prometheus/rules/service-alerts.yml
  • infra/observability/prometheus/prometheus.yml
  • infra/observability/blackbox/blackbox.yml
  • infra/observability/docker-compose.yml
  • rg "ladirecta.com.ar|ladirecta.ar|EcommerceLaDirecta" infra docs: confirma targets, alertas, dashboard y documentacion.
  • git diff --check: sin errores; solo warnings LF/CRLF de Windows.

Deploy ejecutado

Procedimiento documentado disponible en infra/observability/README.md.

Aplicado sobre OpenClaw VPS srv977009:

  1. Se copio infra/observability/ a staging remoto.
  2. Se sincronizo sobre /opt/stacks/observability, preservando .env y secrets/.
  3. Se valido docker compose config.
  4. Se ejecuto docker compose up -d.
  5. Se valido promtool check config.
  6. Se valido promtool check rules.
  7. Se validaron health/readiness y queries.

Servicios observados:

  • obs-prometheus: running.
  • obs-grafana: running.
  • obs-blackbox-exporter: running.
  • obs-alertmanager: running.
  • obs-thanos-query, obs-thanos-sidecar, obs-thanos-storegateway, obs-thanos-compactor: running.
  • obs-node-exporter y obs-cadvisor: existentes del stack O4, sin cambios.

Validaciones remotas:

  • promtool check config /etc/prometheus/prometheus.yml: SUCCESS.
  • promtool check rules /etc/prometheus/rules/*.yml: SUCCESS.
  • Prometheus readiness: Prometheus Server is Ready.
  • Grafana health: database: ok, version 13.0.1+security-01.
  • Blackbox health: Healthy.
  • Alertmanager readiness: OK.
  • Dashboard remoto presente: /opt/stacks/observability/grafana/dashboards/10-executive-board/o97-ecommerce-ladirecta.json.

Queries probadas:

  • probe_success{tenant="ladirecta", service="ecommerce-ladirecta"}: 1 para https://ladirecta.com.ar/ y https://ladirecta.ar/.
  • Uptime 24h: 100 para ambos endpoints.
  • probe_duration_seconds: aprox. 0.69s para https://ladirecta.com.ar/ y 1.04s para https://ladirecta.ar/.
  • SSL days remaining: aprox. 89.77 dias para ambos endpoints.
  • probe_http_status_code: 200 para ambos endpoints.
  • ALERTS{tenant="ladirecta", alertstate="firing"}: sin alertas activas.

Restricciones respetadas

  • No tocar VPS WooCommerce.
  • No abrir puertos.
  • No tocar DNS.
  • No instalar exporters.
  • No tocar WordPress/WooCommerce.
  • No exponer Prometheus ni Alertmanager publicamente.
  • Solo mejorar dashboards, alertas visibles y documentacion en OpenClaw.

Resultado esperado

El operador debe poder abrir Grafana y ver Ecommerce La Directa en la carpeta 10 - Executive Board, con estado, latencia, SSL, uptime y alertas activas de ambos dominios.