Add cluster scanner metrics in metrics-v3 (#19517)

endpoint: /minio/metrics/v3/cluster/scanner
metrics:
 - bucket_scans_finished (counter)
 - bucket_scans_started (counter)
 - directories_scanned (counter)
 - last_activity_nano_seconds (gauge)
 - objects_scanned (counter)
 - versions_scanned (counter)

cmd/metrics-v3-scanner.go

@@ -0,0 +1,66 @@
+// Copyright (c) 2015-2024 MinIO, Inc.
+//
+// This file is part of MinIO Object Storage stack
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+package cmd
+
+import (
+	"context"
+	"time"
+)
+
+const (
+	scannerBucketScansFinished = "bucket_scans_finished"
+	scannerBucketScansStarted  = "bucket_scans_started"
+	scannerDirectoriesScanned  = "directories_scanned"
+	scannerObjectsScanned      = "objects_scanned"
+	scannerVersionsScanned     = "versions_scanned"
+	scannerLastActivitySeconds = "last_activity_seconds"
+)
+
+var (
+	scannerBucketScansFinishedMD = NewCounterMD(scannerBucketScansFinished,
+		"Total number of bucket scans finished since server start")
+	scannerBucketScansStartedMD = NewCounterMD(scannerBucketScansStarted,
+		"Total number of bucket scans started since server start")
+	scannerDirectoriesScannedMD = NewCounterMD(scannerDirectoriesScanned,
+		"Total number of directories scanned since server start")
+	scannerObjectsScannedMD = NewCounterMD(scannerObjectsScanned,
+		"Total number of unique objects scanned since server start")
+	scannerVersionsScannedMD = NewCounterMD(scannerVersionsScanned,
+		"Total number of object versions scanned since server start")
+	scannerLastActivitySecondsMD = NewGaugeMD(scannerLastActivitySeconds,
+		"Time elapsed (in seconds) since last scan activity.")
+)
+
+// loadClusterScannerMetrics - `MetricsLoaderFn` for cluster webhook
+// such as failed objects and directories scanned.
+func loadClusterScannerMetrics(ctx context.Context, m MetricValues, c *metricsCache) error {
+	m.Set(scannerBucketScansFinished, float64(globalScannerMetrics.lifetime(scannerMetricScanBucketDrive)))
+	m.Set(scannerBucketScansStarted, float64(globalScannerMetrics.lifetime(scannerMetricScanBucketDrive)+uint64(globalScannerMetrics.activeDrives())))
+	m.Set(scannerDirectoriesScanned, float64(globalScannerMetrics.lifetime(scannerMetricScanFolder)))
+	m.Set(scannerObjectsScanned, float64(globalScannerMetrics.lifetime(scannerMetricScanObject)))
+	m.Set(scannerVersionsScanned, float64(globalScannerMetrics.lifetime(scannerMetricApplyVersion)))
+
+	dui, err := c.dataUsageInfo.Get()
+	if err != nil {
+		metricsLogIf(ctx, err)
+	} else {
+		m.Set(scannerLastActivitySeconds, time.Since(dui.LastUpdate).Seconds())
+	}
+
+	return nil
+}

cmd/metrics-v3.go

@@ -58,6 +58,7 @@ const (
 	loggerWebhookCollectorPath collectorPath = "/logger/webhook"
 	replicationCollectorPath   collectorPath = "/replication"
 	notificationCollectorPath  collectorPath = "/notification"
+	scannerCollectorPath       collectorPath = "/scanner"
 )
 
 const (
@@ -351,6 +352,18 @@ func newMetricGroups(r *prometheus.Registry) *metricsV3Collection {
 		loadClusterConfigMetrics,
 	)
 
+	scannerMG := NewMetricsGroup(scannerCollectorPath,
+		[]MetricDescriptor{
+			scannerBucketScansFinishedMD,
+			scannerBucketScansStartedMD,
+			scannerDirectoriesScannedMD,
+			scannerObjectsScannedMD,
+			scannerVersionsScannedMD,
+			scannerLastActivitySecondsMD,
+		},
+		loadClusterScannerMetrics,
+	)
+
 	loggerWebhookMG := NewMetricsGroup(loggerWebhookCollectorPath,
 		[]MetricDescriptor{
 			webhookFailedMessagesMD,
@@ -389,6 +402,7 @@ func newMetricGroups(r *prometheus.Registry) *metricsV3Collection {
 		clusterReplicationMG,
 		clusterConfigMG,
 
+		scannerMG,
 		auditMG,
 		loggerWebhookMG,
 	}

docs/metrics/v3.md

@@ -58,6 +58,14 @@ These are metrics about the minio notification functionality
 |----------|------------------------------------------------------|
 | `/notification` | Metrics related to notification functionality |
 
+### Scanner metrics
+
+These are metrics about the minio scanner
+
+| Path       | Description                          |
+|------------|--------------------------------------|
+| `/scanner` | Metrics related to the MinIO scanner |
+
 ### System metrics
 
 These are metrics about the minio process and the node.
@@ -358,3 +366,14 @@ The standard metrics group for GoCollector is not shown below.
 | `minio_notification_events_errors_total`      | `counter` | Events that were failed to be sent to the targets                                        | `server` |
 | `minio_notification_events_sent_total`        | `counter` | Total number of events sent to the targets                                               | `server` |
 | `minio_notification_events_skipped_total`     | `counter` | Events that were skipped to be sent to the targets due to the in-memory queue being full | `server` |
+
+### `/scanner`
+
+| Name                                       | Type      | Help                                                       | Labels   |
+|--------------------------------------------|-----------|------------------------------------------------------------|----------|
+| `minio_scanner_bucket_scans_finished`      | `counter` | Total number of bucket scans finished since server start   | `server` |
+| `minio_scanner_bucket_scans_started`       | `counter` | Total number of bucket scans started since server start    | `server` |
+| `minio_scanner_directories_scanned`        | `counter` | Total number of directories scanned since server start     | `server` |
+| `minio_scanner_last_activity_seconds`      | `gauge`   | Time elapsed (in seconds) since last scan activity         | `server` |
+| `minio_scanner_objects_scanned`            | `counter` | Total number of unique objects scanned since server start  | `server` |
+| `minio_scanner_versions_scanned`           | `counter` | Total number of object versions scanned since server start | `server` |