dns: add enhanced metrics (#1999) (#2040)

Co-authored-by: Matthew Wimpelberg <120263653+mwimpelberg28@users.noreply.github.com>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

cmd/windows_exporter/0_service.go

@@ -38,7 +38,7 @@ var (
 	stopCh = make(chan struct{})
 
 	// serviceManagerFinishedCh is a channel to send a signal to the main function that the service manager has stopped the service.
-	serviceManagerFinishedCh = make(chan struct{})
+	serviceManagerFinishedCh = make(chan struct{}, 1)
 )
 
 // IsService variable declaration allows initiating time-sensitive components like registering the Windows service

docs/collector.dns.md

@@ -3,14 +3,19 @@
 The dns collector exposes metrics about the DNS server
 
 |||
--|-
-Metric name prefix  | `dns`
-Classes             | [`Win32_PerfRawData_DNS_DNS`](https://technet.microsoft.com/en-us/library/cc977686.aspx)
-Enabled by default? | No
+-|-|-
+Metric name prefix  | `dns` |
+Classes             | [`Win32_PerfRawData_DNS_DNS`](https://technet.microsoft.com/en-us/library/cc977686.aspx) |
+Enabled by default | Yes |
+Metric name prefix (error stats) | `windows_dns` |
+Classes             | [`MicrosoftDNS_Statistic`](https://learn.microsoft.com/en-us/windows/win32/dns/dns-wmi-provider-overview) |
+Enabled by default (error stats)? | Yes |
 
 ## Flags
 
-None
+Name | Description
+-----|------------
+`collector.dns.enabled` | Comma-separated list of collectors to use. Available collectors: `metrics`, `error_stats`. Defaults to all collectors if not specified.
 
 ## Metrics
 
@@ -38,12 +43,56 @@ Name | Description | Type | Labels
 `windows_dns_wins_queries_total` | _Not yet documented_ | counter | `direction`
 `windows_dns_wins_responses_total` | _Not yet documented_ | counter | `direction`
 `windows_dns_unmatched_responses_total` | _Not yet documented_ | counter | None
+`windows_dns_error_stats_total` | DNS error statistics from MicrosoftDNS_Statistic | counter | `name`, `collection_name`, `dns_server`
+
+### Sub-collectors
+
+The DNS collector is split into two sub-collectors:
+
+1. `metrics` - Collects standard DNS performance metrics using PDH (Performance Data Helper)
+2. `wmi_stats` - Collects DNS error statistics from the MicrosoftDNS_Statistic WMI class
+
+By default, both sub-collectors are enabled. You can enable specific sub-collectors using the `collector.dns.enabled` flag.
+
+### Example Usage
+
+To enable only DNS error statistics collection:
+```powershell
+windows_exporter.exe --collector.dns.enabled=wmi_stats
+```
+
+To enable only standard DNS metrics:
+```powershell
+windows_exporter.exe --collector.dns.enabled=metrics
+```
+
+To enable both (default behavior):
+```powershell
+windows_exporter.exe --collector.dns.enabled=metrics,wmi_stats
+```
 
 ### Example metric
-_This collector does not yet have explained examples, we would appreciate your help adding them!_
+```
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="BadKey"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="BadSig"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="BadTime"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="FormError"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="Max"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NoError"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NotAuth"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NotImpl"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NotZone"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NxDomain"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="NxRRSet"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="Refused"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="ServFail"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="UnknownError"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="YxDomain"} 0
+windows_dns_wmi_stats_total{collection_name="Error Stats",dns_server="EC2AMAZ-5NNM8M1",name="YxRRSet"} 0
+```
 
 ## Useful queries
 _This collector does not yet have any useful queries added, we would appreciate your help adding them!_
 
 ## Alerting examples
-_This collector does not yet have alerting examples, we would appreciate your help adding them!_
+_This collector does not yet have alerting examples, we would appreciate your help adding them!_

docs/collector.exchange.md

@@ -43,7 +43,7 @@ Comma-separated list of collectors to use, for example: `--collectors.exchange.e
 | `windows_exchange_transport_queues_messages_submitted_total`                | Messages Submitted Total                                                                                    |
 | `windows_exchange_transport_queues_messages_delayed_total`                  | Messages Delayed Total                                                                                      |
 | `windows_exchange_transport_queues_messages_completed_delivery_total`       | Messages Completed Delivery Total                                                                           |
-| `windows_exchange_transport_queues_shadow_queue_length`                     | Shadow Queue Length                                                                                         |
+| `windows_exchange_transport_queues_aggregate_shadow_queue_length`           | The current number of messages in shadow queues                                                             |
 | `windows_exchange_transport_queues_submission_queue_length`                 | Submission Queue Length                                                                                     |
 | `windows_exchange_transport_queues_delay_queue_length`                      | Delay Queue Length                                                                                          |
 | `windows_exchange_transport_queues_items_completed_delivery_total`          | Items Completed Delivery Total                                                                              |
@@ -54,7 +54,7 @@ Comma-separated list of collectors to use, for example: `--collectors.exchange.e
 | `windows_exchange_http_proxy_avg_auth_latency`                              | Average time spent authenticating CAS requests over the last 200 samples                                    |
 | `windows_exchange_http_proxy_outstanding_proxy_requests`                    | Number of concurrent outstanding proxy requests                                                             |
 | `windows_exchange_http_proxy_requests_total`                                | Number of proxy requests processed each second                                                              |
-| `windows_exchange_avail_service_requests_per_sec`                           | Number of requests serviced per second                                                                      |
+| `windows_exchange_availability_service_requests_per_sec`                    | Number of requests serviced per second                                                                      |
 | `windows_exchange_owa_current_unique_users`                                 | Number of unique users currently logged on to Outlook Web App                                               |
 | `windows_exchange_owa_requests_total`                                       | Number of requests handled by Outlook Web App per second                                                    |
 | `windows_exchange_autodiscover_requests_total`                              | Number of autodiscover service requests processed each second                                               |
@@ -77,4 +77,3 @@ _This collector does not yet have any useful queries added, we would appreciate
 
 ## Alerting examples
 _This collector does not yet have alerting examples, we would appreciate your help adding them!_
-

internal/collector/dns/dns.go

@@ -16,8 +16,11 @@
 package dns
 
 import (
+	"errors"
 	"fmt"
 	"log/slog"
+	"slices"
+	"strings"
 
 	"github.com/alecthomas/kingpin/v2"
 	"github.com/prometheus-community/windows_exporter/internal/mi"
@@ -26,12 +29,23 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 )
 
-const Name = "dns"
+const (
+	Name                 = "dns"
+	subCollectorMetrics  = "metrics"
+	subCollectorWMIStats = "wmi_stats"
+)
 
-type Config struct{}
+type Config struct {
+	CollectorsEnabled []string `yaml:"collectors_enabled"`
+}
 
 //nolint:gochecknoglobals
-var ConfigDefaults = Config{}
+var ConfigDefaults = Config{
+	CollectorsEnabled: []string{
+		subCollectorMetrics,
+		subCollectorWMIStats,
+	},
+}
 
 // A Collector is a Prometheus Collector for WMI Win32_PerfRawData_DNS_DNS metrics.
 type Collector struct {
@@ -40,6 +54,9 @@ type Collector struct {
 	perfDataCollector *pdh.Collector
 	perfDataObject    []perfDataCounterValues
 
+	miSession *mi.Session
+	miQuery   mi.Query
+
 	dynamicUpdatesFailures        *prometheus.Desc
 	dynamicUpdatesQueued          *prometheus.Desc
 	dynamicUpdatesReceived        *prometheus.Desc
@@ -62,6 +79,7 @@ type Collector struct {
 	zoneTransferResponsesReceived *prometheus.Desc
 	zoneTransferSuccessReceived   *prometheus.Desc
 	zoneTransferSuccessSent       *prometheus.Desc
+	dnsWMIStats                   *prometheus.Desc
 }
 
 func New(config *Config) *Collector {
@@ -69,6 +87,10 @@ func New(config *Config) *Collector {
 		config = &ConfigDefaults
 	}
 
+	if config.CollectorsEnabled == nil {
+		config.CollectorsEnabled = ConfigDefaults.CollectorsEnabled
+	}
+
 	c := &Collector{
 		config: *config,
 	}
@@ -76,8 +98,26 @@ func New(config *Config) *Collector {
 	return c
 }
 
-func NewWithFlags(_ *kingpin.Application) *Collector {
-	return &Collector{}
+func NewWithFlags(app *kingpin.Application) *Collector {
+	c := &Collector{
+		config: ConfigDefaults,
+	}
+	c.config.CollectorsEnabled = make([]string, 0)
+
+	var collectorsEnabled string
+
+	app.Flag(
+		"collector.dns.enabled",
+		"Comma-separated list of collectors to use. Defaults to all, if not specified.",
+	).Default(strings.Join(ConfigDefaults.CollectorsEnabled, ",")).StringVar(&collectorsEnabled)
+
+	app.Action(func(*kingpin.ParseContext) error {
+		c.config.CollectorsEnabled = strings.Split(collectorsEnabled, ",")
+
+		return nil
+	})
+
+	return c
 }
 
 func (c *Collector) GetName() string {
@@ -90,7 +130,31 @@ func (c *Collector) Close() error {
 	return nil
 }
 
-func (c *Collector) Build(_ *slog.Logger, _ *mi.Session) error {
+func (c *Collector) Build(_ *slog.Logger, miSession *mi.Session) error {
+	for _, collector := range c.config.CollectorsEnabled {
+		if !slices.Contains([]string{subCollectorMetrics, subCollectorWMIStats}, collector) {
+			return fmt.Errorf("unknown sub collector: %s. Possible values: %s", collector,
+				strings.Join([]string{subCollectorMetrics, subCollectorWMIStats}, ", "),
+			)
+		}
+	}
+
+	if slices.Contains(c.config.CollectorsEnabled, subCollectorMetrics) {
+		if err := c.buildMetricsCollector(); err != nil {
+			return err
+		}
+	}
+
+	if slices.Contains(c.config.CollectorsEnabled, subCollectorWMIStats) {
+		if err := c.buildErrorStatsCollector(miSession); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (c *Collector) buildMetricsCollector() error {
 	c.zoneTransferRequestsReceived = prometheus.NewDesc(
 		prometheus.BuildFQName(types.Namespace, Name, "zone_transfer_requests_received_total"),
 		"Number of zone transfer requests (AXFR/IXFR) received by the master DNS server",
@@ -224,6 +288,13 @@ func (c *Collector) Build(_ *slog.Logger, _ *mi.Session) error {
 		nil,
 	)
 
+	c.dnsWMIStats = prometheus.NewDesc(
+		prometheus.BuildFQName(types.Namespace, Name, "wmi_stats_total"),
+		"DNS WMI statistics from MicrosoftDNS_Statistic",
+		[]string{"name", "collection_name", "dns_server"},
+		nil,
+	)
+
 	var err error
 
 	c.perfDataCollector, err = pdh.NewCollector[perfDataCounterValues](pdh.CounterTypeRaw, "DNS", pdh.InstancesAll)
@@ -234,9 +305,43 @@ func (c *Collector) Build(_ *slog.Logger, _ *mi.Session) error {
 	return nil
 }
 
+func (c *Collector) buildErrorStatsCollector(miSession *mi.Session) error {
+	if miSession == nil {
+		return errors.New("miSession is nil")
+	}
+
+	query, err := mi.NewQuery("SELECT Name, CollectionName, Value, DnsServerName FROM MicrosoftDNS_Statistic WHERE CollectionName = 'Error Stats'")
+	if err != nil {
+		return fmt.Errorf("failed to create query: %w", err)
+	}
+
+	c.miSession = miSession
+	c.miQuery = query
+
+	return nil
+}
+
 // Collect sends the metric values for each metric
 // to the provided prometheus Metric channel.
 func (c *Collector) Collect(ch chan<- prometheus.Metric) error {
+	errs := make([]error, 0)
+
+	if slices.Contains(c.config.CollectorsEnabled, subCollectorMetrics) {
+		if err := c.collectMetrics(ch); err != nil {
+			errs = append(errs, fmt.Errorf("failed collecting metrics: %w", err))
+		}
+	}
+
+	if slices.Contains(c.config.CollectorsEnabled, subCollectorWMIStats) {
+		if err := c.collectErrorStats(ch); err != nil {
+			errs = append(errs, fmt.Errorf("failed collecting WMI statistics: %w", err))
+		}
+	}
+
+	return errors.Join(errs...)
+}
+
+func (c *Collector) collectMetrics(ch chan<- prometheus.Metric) error {
 	err := c.perfDataCollector.Collect(&c.perfDataObject)
 	if err != nil {
 		return fmt.Errorf("failed to collect DNS metrics: %w", err)
@@ -493,3 +598,24 @@ func (c *Collector) Collect(ch chan<- prometheus.Metric) error {
 
 	return nil
 }
+
+func (c *Collector) collectErrorStats(ch chan<- prometheus.Metric) error {
+	var stats []Statistic
+	if err := c.miSession.Query(&stats, mi.NamespaceRootMicrosoftDNS, c.miQuery); err != nil {
+		return fmt.Errorf("failed to query DNS statistics: %w", err)
+	}
+
+	// Collect DNS error statistics
+	for _, stat := range stats {
+		ch <- prometheus.MustNewConstMetric(
+			c.dnsWMIStats,
+			prometheus.CounterValue,
+			float64(stat.Value),
+			stat.Name,
+			stat.CollectionName,
+			stat.DnsServerName,
+		)
+	}
+
+	return nil
+}

internal/collector/dns/types.go

@@ -105,3 +105,11 @@ type perfDataCounterValues struct {
 	_                              float64 `perfdata:"Zone Transfer SOA Request Sent"`
 	_                              float64 `perfdata:"Zone Transfer Success"`
 }
+
+// Statistic represents the structure for DNS error statistics
+type Statistic struct {
+	Name           string `mi:"Name"`
+	CollectionName string `mi:"CollectionName"`
+	Value          uint64 `mi:"Value"`
+	DnsServerName  string `mi:"DnsServerName"`
+}

internal/collector/exchange/exchange_autodiscover.go

@@ -37,7 +37,7 @@ type perfDataCounterValuesAutoDiscover struct {
 func (c *Collector) buildAutoDiscover() error {
 	var err error
 
-	c.perfDataCollectorAutoDiscover, err = pdh.NewCollector[perfDataCounterValuesAutoDiscover](pdh.CounterTypeRaw, "MSExchange Autodiscover", pdh.InstancesAll)
+	c.perfDataCollectorAutoDiscover, err = pdh.NewCollector[perfDataCounterValuesAutoDiscover](pdh.CounterTypeRaw, "MSExchangeAutodiscover", nil)
 	if err != nil {
 		return fmt.Errorf("failed to create MSExchange Autodiscover collector: %w", err)
 	}

internal/collector/exchange/exchange_availability_service.go

@@ -31,7 +31,7 @@ type collectorAvailabilityService struct {
 }
 
 type perfDataCounterValuesAvailabilityService struct {
-	RequestsPerSec float64 `perfdata:"Requests/sec"`
+	AvailabilityRequestsPerSec float64 `perfdata:"Availability Requests (sec)"`
 }
 
 func (c *Collector) buildAvailabilityService() error {
@@ -43,7 +43,7 @@ func (c *Collector) buildAvailabilityService() error {
 	}
 
 	c.availabilityRequestsSec = prometheus.NewDesc(
-		prometheus.BuildFQName(types.Namespace, Name, "avail_service_requests_per_sec"),
+		prometheus.BuildFQName(types.Namespace, Name, "availability_service_requests_per_sec"),
 		"Number of requests serviced per second",
 		nil,
 		nil,
@@ -62,7 +62,7 @@ func (c *Collector) collectAvailabilityService(ch chan<- prometheus.Metric) erro
 		ch <- prometheus.MustNewConstMetric(
 			c.availabilityRequestsSec,
 			prometheus.CounterValue,
-			data.RequestsPerSec,
+			data.AvailabilityRequestsPerSec,
 		)
 	}
 

internal/collector/exchange/exchange_transport_queues.go

@@ -39,7 +39,7 @@ type collectorTransportQueues struct {
 	messagesSubmittedTotal                  *prometheus.Desc
 	messagesDelayedTotal                    *prometheus.Desc
 	messagesCompletedDeliveryTotal          *prometheus.Desc
-	shadowQueueLength                       *prometheus.Desc
+	aggregateShadowQueueLength              *prometheus.Desc
 	submissionQueueLength                   *prometheus.Desc
 	delayQueueLength                        *prometheus.Desc
 	itemsCompletedDeliveryTotal             *prometheus.Desc
@@ -63,7 +63,7 @@ type perfDataCounterValuesTransportQueues struct {
 	MessagesSubmittedTotal                  float64 `perfdata:"Messages Submitted Total"`
 	MessagesDelayedTotal                    float64 `perfdata:"Messages Delayed Total"`
 	MessagesCompletedDeliveryTotal          float64 `perfdata:"Messages Completed Delivery Total"`
-	ShadowQueueLength                       float64 `perfdata:"Shadow Queue Length"`
+	AggregateShadowQueueLength              float64 `perfdata:"Aggregate Shadow Queue Length"`
 	SubmissionQueueLength                   float64 `perfdata:"Submission Queue Length"`
 	DelayQueueLength                        float64 `perfdata:"Delay Queue Length"`
 	ItemsCompletedDeliveryTotal             float64 `perfdata:"Items Completed Delivery Total"`
@@ -152,9 +152,9 @@ func (c *Collector) buildTransportQueues() error {
 		[]string{"name"},
 		nil,
 	)
-	c.shadowQueueLength = prometheus.NewDesc(
-		prometheus.BuildFQName(types.Namespace, Name, "transport_queues_shadow_queue_length"),
-		"Shadow Queue Length",
+	c.aggregateShadowQueueLength = prometheus.NewDesc(
+		prometheus.BuildFQName(types.Namespace, Name, "transport_queues_aggregate_shadow_queue_length"),
+		"The current number of messages in shadow queues.",
 		[]string{"name"},
 		nil,
 	)
@@ -280,9 +280,9 @@ func (c *Collector) collectTransportQueues(ch chan<- prometheus.Metric) error {
 			labelName,
 		)
 		ch <- prometheus.MustNewConstMetric(
-			c.shadowQueueLength,
+			c.aggregateShadowQueueLength,
 			prometheus.GaugeValue,
-			data.ShadowQueueLength,
+			data.AggregateShadowQueueLength,
 			labelName,
 		)
 		ch <- prometheus.MustNewConstMetric(

internal/collector/service/service.go

@@ -420,15 +420,6 @@ func (c *Collector) getProcessStartTime(pid uint32) (uint64, error) {
 		return 0, fmt.Errorf("failed to open process %w", err)
 	}
 
-	defer func(handle windows.Handle) {
-		err := windows.CloseHandle(handle)
-		if err != nil {
-			c.logger.Warn("failed to close process handle",
-				slog.Any("err", err),
-			)
-		}
-	}(handle)
-
 	var (
 		creation windows.Filetime
 		exit     windows.Filetime
@@ -437,6 +428,14 @@ func (c *Collector) getProcessStartTime(pid uint32) (uint64, error) {
 	)
 
 	err = windows.GetProcessTimes(handle, &creation, &exit, &krn, &user)
+
+	if err := windows.CloseHandle(handle); err != nil {
+		c.logger.LogAttrs(context.Background(), slog.LevelWarn, "failed to close process handle",
+			slog.Any("err", err),
+			slog.Uint64("pid", uint64(pid)),
+		)
+	}
+
 	if err != nil {
 		return 0, fmt.Errorf("failed to get process times %w", err)
 	}
@@ -477,7 +476,7 @@ func (c *Collector) getServiceConfig(service *mgr.Service) (mgr.Config, error) {
 		*buf = make([]byte, bytesNeeded)
 	}
 
-	c.serviceConfigPoolBytes.Put(buf)
+	defer c.serviceConfigPoolBytes.Put(buf)
 
 	return mgr.Config{
 		BinaryPathName:   windows.UTF16PtrToString(serviceConfig.BinaryPathName),

internal/collector/terminal_services/terminal_services.go

@@ -437,7 +437,7 @@ func (c *Collector) collectWTSSessions(ch chan<- prometheus.Metric) error {
 	for _, session := range sessions {
 		// only connect metrics for remote named sessions
 		n := strings.ReplaceAll(session.SessionName, "#", " ")
-		if n == "" || n == "Services" {
+		if n == "Services" {
 			continue
 		}
 

internal/mi/types.go

@@ -51,6 +51,7 @@ var (
 	NamespaceRootWindowsFSRM       = utils.Must(NewNamespace("root/microsoft/windows/fsrm"))
 	NamespaceRootWebAdministration = utils.Must(NewNamespace("root/WebAdministration"))
 	NamespaceRootMSCluster         = utils.Must(NewNamespace("root/MSCluster"))
+	NamespaceRootMicrosoftDNS      = utils.Must(NewNamespace("root/MicrosoftDNS"))
 )
 
 type Query *uint16