config.alloy

remotecfg {
  url            = "..."
  id             = "..."
  poll_frequency = "60s"

  basic_auth {
    username = "..."
    password = "..."
  }
}

prometheus.remote_write "metrics_service" {
  endpoint {
    url = "..."

    basic_auth {
      username = "..."
      password = "..."
    }

    write_relabel_config {
      action        = "drop"
      source_labels = ["__name__"]
      regex         = "(alloy_component_.*|node_cooling_.*|node_cpu_frequency_.*|node_cpu_scaling_.*|node_disk_.*|node_scrape_collector.*|prometheus_remote_storage_.*)"
    }
  }
}

loki.write "grafana_cloud_loki" {
  endpoint {
    url = "..."

    basic_auth {
      username = "..."
      password = "..."
    }
  }
}

/////////////////////////////////////////////////////////////////////////////
//                         Firefleet node metrics                          //
/////////////////////////////////////////////////////////////////////////////

// Add standard labels to metrics coming from node_exporter
discovery.relabel "node_exporter" {
  targets = prometheus.exporter.unix.node_exporter.targets

  rule {
    target_label = "instance"
    replacement  = constants.hostname
  }

  rule {
    target_label = "job"
    replacement  = "integrations/node_exporter"
  }
}

// Collect system metrics
prometheus.exporter.unix "node_exporter" {
  // Disable unnecessary collectors to reduce overhead (and fit the free tier)
  disable_collectors = [
    "arp",
    "bcache",
    "bonding",
    "btrfs",
    "buddyinfo",
    "dmi",
    "edac",
    "entropy",
    "fibrechannel",
    "hwmon",
    "infiniband",
    "ipvs",
    "mdadm",
    "netstat",
    "pressure",
    "schedstat",
    "sockstat",
    "softnet",
    "timex",
    "udp_queues",
    "uname",
    "xfs",
    "zfs",
  ]
  enable_collectors = ["meminfo"]

  filesystem {
    // Exclude filesystem types that aren't relevant for monitoring
    fs_types_exclude     = "^(autofs|binfmt_misc|bpf|cgroup2?|configfs|debugfs|devpts|devtmpfs|tmpfs|fusectl|hugetlbfs|iso9660|mqueue|nsfs|overlay|proc|procfs|pstore|rpc_pipefs|securityfs|selinuxfs|squashfs|sysfs|tracefs)$"
    
    // Exclude mount points that aren't relevant for monitoring
    mount_points_exclude = "^/(dev|proc|run/credentials/.+|sys)($|/)"
    
    mount_timeout        = "5s"
  }

  netclass {
    // Ignore virtual and container network interfaces
    ignored_devices = "^(veth.*|cali.*|docker.*|play.*|[a-f0-9]{15})$"
  }

  netdev {
    // Exclude virtual and container network interfaces from device metrics
    device_exclude = "^(veth.*|cali.*|docker.*|play.*|[a-f0-9]{15})$"
  }
}

// Scrape metrics from the node_exporter
prometheus.scrape "node_exporter" {
  scrape_interval = "60s"
  
  targets    = discovery.relabel.node_exporter.output
  
  forward_to = [prometheus.remote_write.metrics_service.receiver]
}

/////////////////////////////////////////////////////////////////////////////
//                          Firefleet node logs                            //
/////////////////////////////////////////////////////////////////////////////

// Convert well-known journal message fields into labels
discovery.relabel "journal" {
  targets = []

  rule {
    source_labels = ["__journal__systemd_unit"]
    target_label  = "unit"
  }

  rule {
    source_labels = ["__journal__boot_id"]
    target_label  = "boot_id"
  }

  rule {
    source_labels = ["__journal__transport"]
    target_label  = "transport"
  }

  rule {
    source_labels = ["__journal_priority_keyword"]
    target_label  = "level"
  }
}

// journald logs
loki.source.journal "journal" {
  // Only collect logs from the last 2 hours
  max_age       = "2h0m0s"

  labels        = { source = "journal", instance = constants.hostname }
  
  relabel_rules = discovery.relabel.journal.rules
  
  forward_to    = [loki.write.grafana_cloud_loki.receiver]
}

// system log files
local.file_match "system_logs" {
  path_targets = [{
    // Target localhost for log collection
    __address__ = "localhost",

    // Collect standard system logs
    __path__    = "/var/log/{syslog,messages,*.log}",

    instance    = constants.hostname,
    
    job         = "integrations/node_exporter",
  }]
}

// system logs (via file scraping)
loki.source.file "system_logs" {
  targets    = local.file_match.system_logs.targets
  
  forward_to = [loki.write.grafana_cloud_loki.receiver]
}


/////////////////////////////////////////////////////////////////////////////
//                        Firefleet Docker logs                            //
/////////////////////////////////////////////////////////////////////////////

discovery.docker "docker_containers" {
  host             = "unix:///var/run/docker.sock"
  refresh_interval = "5s"
}

discovery.relabel "docker_containers" {
  targets = []

  rule {
    target_label = "job"
    replacement  = "integrations/docker"
  }

  rule {
    target_label = "instance"
    replacement  = constants.hostname
  }

  rule {
    source_labels = ["__meta_docker_container_name"]
    regex         = "/(.*)"
    target_label  = "container"
  }

  rule {
    source_labels = ["__meta_docker_container_log_stream"]
    target_label  = "stream"
  }
}

loki.source.docker "docker_containers" {
  host            = "unix:///var/run/docker.sock"
  targets         = discovery.docker.docker_containers.targets
  labels          = { source = "docker" }
  relabel_rules   = discovery.relabel.docker_containers.rules
  forward_to      = [loki.write.grafana_cloud_loki.receiver]
  refresh_interval = "5s"
}