dns collector bug windows exporter Alloy 1.9.2

[bbtango]

What's wrong?

adding the text_file collector result in a bug where alloy is trying to enable dns collector even if this one isn't included in the config :

Error: C:\Program Files\GrafanaLabs\Alloy\config.alloy:5:1: Failed to build component: building component: error build collector dns: unknown sub collector: . Possible values: metrics, wmi_stats

 4 |   
 5 |   prometheus.exporter.windows "default" {
   |  _^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 6 | |   enabled_collectors = ["text_file", "cpu", "logical_disk", "net", "service","system", "memory", "iis", "mssql", "scheduled_task", "process", "tcp"]
 7 | | 
 8 | |   logical_disk {
 9 | |     exclude = "test"
10 | |   }
11 | |   network {
12 | |     exclude = "test"
13 | |   }
14 | |   process {
15 | |     exclude = "test"
16 | |   }
17 | |   service {
18 | |     exclude = "test"
19 | |   }
20 | |   iis {
21 | |     app_exclude = "test"
22 | |   }
23 | |   mssql {
24 | |     enabled_classes = ["bufman","databases","genstats","info","locks","memmgr","sqlerrors","sqlstats","transactions","waitstats"]
25 | |   }
26 | |   scheduled_task {
27 | |     exclude = "test"
28 | |   }
29 | |   tcp {
30 | |     enabled_list = ["metrics","connections_state"]
31 | |   }
32 | | }
   | |_^
33 |

Steps to reproduce

Use this prometheus.exporter.windows config :

prometheus.exporter.windows "default" {
  enabled_collectors = ["text_file", "cpu", "logical_disk", "net", "service","system", "memory", "iis", "mssql", "scheduled_task", "process", "tcp"]

  logical_disk {
    exclude = "test"
  }
  network {
    exclude = "test"
  }
  process {
    exclude = "test"
  }
  service {
    exclude = "test"
  }
  iis {
    app_exclude = "test"
  }
  mssql {
    enabled_classes = ["bufman","databases","genstats","info","locks","memmgr","sqlerrors","sqlstats","transactions","waitstats"]
  }
  scheduled_task {
    exclude = "test"
  }
  tcp {
    enabled_list = ["metrics","connections_state"]
  }
}

System information

Windows 10 Pro 22H2

Software version

Alloy 1.9.2

Configuration

// ####################################
// Windows Server Metrics Configuration
// ####################################

prometheus.exporter.windows "default" {
  enabled_collectors = ["text_file", "cpu", "logical_disk", "net", "service","system", "memory", "iis", "mssql", "scheduled_task", "process", "tcp"]

  logical_disk {
    exclude = "test"
  }
  network {
    exclude = "test"
  }
  process {
    exclude = "test"
  }
  service {
    exclude = "test"
  }
  iis {
    app_exclude = "test"
  }
  mssql {
    enabled_classes = ["bufman","databases","genstats","info","locks","memmgr","sqlerrors","sqlstats","transactions","waitstats"]
  }
  scheduled_task {
    exclude = "test"
  }
  tcp {
    enabled_list = ["metrics","connections_state"]
  }
}

prometheus.remote_write "send_to_prom" {
  endpoint {
    url = "http:/adresse:9090/api/v1/write"
  }
}

prometheus.scrape "default" {
  targets = prometheus.exporter.windows.default.targets
  forward_to = [prometheus.remote_write.send_to_prom.receiver]
}

// ####################################
// Windows Server Logs Configuration
// ####################################

loki.source.windowsevent "task_scheduler_events" {
  eventlog_name = "Microsoft-Windows-TaskScheduler/Operational"
  forward_to    = [loki.process.endpoint.receiver]
}

loki.source.windowsevent "system_events" {
  eventlog_name = "System"
  use_incoming_timestamp = true
  forward_to    = [loki.process.endpoint.receiver]
}

loki.process "endpoint" {
  forward_to = [loki.write.endpoint.receiver]
    stage.json {
      expressions = {
        message = "",
        Overwritten = "",
        source = "",
        computer = "",
        level = "",
        event_id = "",
        eventRecordID = "",
        task = "",
        channel = "",
        componenet_id = "",
        execution_processId = "execution.processId",
        execution_processName = "execution.processName",
      }
    }

    stage.structured_metadata {
      values = {
        "eventRecordID" = "",
        "channel" = "",
        "event_id" = "",
        "task" = "",
        "level" = "",
        "componenet_id" = "",
        "execution_processId" = "",
        "execution_processName" = "",
      }
    }

    stage.eventlogmessage {
      source = "message"
      overwrite_existing = true
    }

    stage.labels {
      values = {
        "service_name" = "source",
      }
    }

    stage.output {
      source = "message"
    }
}

loki.write "endpoint" {
  endpoint {
    url = "http:/adresse:3100/loki/api/v1/push"
  }
}

livedebugging {
  enabled = true
}

Logs

[dehaansa]

Interesting, I wasn't able to reproduce this initially, I can take a deeper look. Are you certain that is the active configuration when experiencing the issue? I've confirmed that the default values for the dns collector are correctly wmi_stats and metrics, but I don't see an overriding configuration in your config. Is there any chance you're somehow sending runtime flags (collector.dns.enabled) that the exporter might interpret as a list of collectors for dns?

[dehaansa]

Turns out there's two issues at play - we aren't propagating the default DNS config, but as it's not being used that should be fine, the main issue is that the collector name is incorrect - it should be textfile not text_file.