monitoring/TODO.md

• Split exporters dans un job dédié (pour pouvoir tourner sur node_pool spécifique)

• Créer monitoring-agent type system avec vector + node-exporter

• images

  • prometheus
  • ping-exporter
  • blackbox-exporter
  • consul-exporter
  • vector
  • loki
  • grafana
  • nomad-vector-logger

• pki roles:

  • monitoring -> prom
  • consul -> prom
  • monitoring -> am

• vault pol

  • prometheus
    • issue prom on monitoring
    • issue prom on consul
  • consul-exporter
    • issue consul-exporter on consul
  • alertmanager
    • issue alertmanager on monitoring

• consul defaults & intentions

  • prometheus
  • alertmanager
  • loki

• tasks

  • alertmanager
  • vector-aggregator
  • vector-agent (dans job agent)
  • loki (modulariser ou laisser en monolithique ?)
  • grafana
  • cluster-metrics (job exporters)

• questions / various

  • prom rules: keep or move to a -conf bundle ?
  • config alertes am (recipient + routing)
  • http and tcp probes, as exporters are now in a dedicated job
  • alertmanager & rules for loki
  • bootstrap grafana