Fix export of stale device error metrics for unmounted filesystems

Instead of maintaining a counter metric for device errors in memory,
this change exports a gauge and uses const metrics to avoid leaking
metrics for unmounted filesystems.

collector/filesystem_common.go

@@ -44,11 +44,11 @@ var (
 )
 
 type filesystemCollector struct {
-	ignoredMountPointsPattern *regexp.Regexp
-	ignoredFSTypesPattern     *regexp.Regexp
-	sizeDesc, freeDesc, availDesc,
-	filesDesc, filesFreeDesc, roDesc *prometheus.Desc
-	devErrors *prometheus.CounterVec
+	ignoredMountPointsPattern     *regexp.Regexp
+	ignoredFSTypesPattern         *regexp.Regexp
+	sizeDesc, freeDesc, availDesc *prometheus.Desc
+	filesDesc, filesFreeDesc      *prometheus.Desc
+	roDesc, deviceErrorDesc       *prometheus.Desc
 }
 
 type filesystemLabels struct {
@@ -56,8 +56,10 @@ type filesystemLabels struct {
 }
 
 type filesystemStats struct {
-	labels                                  filesystemLabels
-	size, free, avail, files, filesFree, ro float64
+	labels            filesystemLabels
+	size, free, avail float64
+	files, filesFree  float64
+	ro, deviceError   float64
 }
 
 func init() {
@@ -106,10 +108,11 @@ func NewFilesystemCollector() (Collector, error) {
 		filesystemLabelNames, nil,
 	)
 
-	devErrors := prometheus.NewCounterVec(prometheus.CounterOpts{
-		Name: prometheus.BuildFQName(Namespace, subsystem, "device_errors_total"),
-		Help: "Total number of errors occurred when getting stats for device",
-	}, filesystemLabelNames)
+	deviceErrorDesc := prometheus.NewDesc(
+		prometheus.BuildFQName(Namespace, subsystem, "device_error"),
+		"Whether an error occured while getting statistics for the given device.",
+		filesystemLabelNames, nil,
+	)
 
 	return &filesystemCollector{
 		ignoredMountPointsPattern: mountPointPattern,
@@ -120,7 +123,7 @@ func NewFilesystemCollector() (Collector, error) {
 		filesDesc:                 filesDesc,
 		filesFreeDesc:             filesFreeDesc,
 		roDesc:                    roDesc,
-		devErrors:                 devErrors,
+		deviceErrorDesc:           deviceErrorDesc,
 	}, nil
 }
 
@@ -137,6 +140,14 @@ func (c *filesystemCollector) Update(ch chan<- prometheus.Metric) error {
 		}
 		seen[s.labels] = true
 
+		ch <- prometheus.MustNewConstMetric(
+			c.deviceErrorDesc, prometheus.GaugeValue,
+			s.deviceError, s.labels.device, s.labels.mountPoint, s.labels.fsType,
+		)
+		if s.deviceError > 0 {
+			continue
+		}
+
 		ch <- prometheus.MustNewConstMetric(
 			c.sizeDesc, prometheus.GaugeValue,
 			s.size, s.labels.device, s.labels.mountPoint, s.labels.fsType,
@@ -162,6 +173,5 @@ func (c *filesystemCollector) Update(ch chan<- prometheus.Metric) error {
 			s.ro, s.labels.device, s.labels.mountPoint, s.labels.fsType,
 		)
 	}
-	c.devErrors.Collect(ch)
 	return nil
 }

collector/filesystem_linux.go

@@ -46,13 +46,15 @@ func (c *filesystemCollector) GetStats() ([]filesystemStats, error) {
 			log.Debugf("Ignoring fs type: %s", labels.fsType)
 			continue
 		}
-		labelValues := []string{labels.device, labels.mountPoint, labels.fsType}
+
 		buf := new(syscall.Statfs_t)
 		err := syscall.Statfs(labels.mountPoint, buf)
 		if err != nil {
-			c.devErrors.WithLabelValues(labelValues...).Inc()
-			log.Debugf("Statfs on %s returned %s",
-				labels.mountPoint, err)
+			stats = append(stats, filesystemStats{
+				labels:      labels,
+				deviceError: 1,
+			})
+			log.Errorf("Error on statfs() system call for %q: %s", labels.mountPoint, err)
 			continue
 		}
 
@@ -82,11 +84,14 @@ func mountPointDetails() ([]filesystemLabels, error) {
 	defer file.Close()
 
 	filesystems := []filesystemLabels{}
-
 	scanner := bufio.NewScanner(file)
 	for scanner.Scan() {
 		parts := strings.Fields(scanner.Text())
-		filesystems = append(filesystems, filesystemLabels{parts[0], parts[1], parts[2]})
+		filesystems = append(filesystems, filesystemLabels{
+			device:     parts[0],
+			mountPoint: parts[1],
+			fsType:     parts[2],
+		})
 	}
 	return filesystems, scanner.Err()
 }