Feature parity between Python disk check and Go disk check

cmd/agent/dist/conf.d/disk.d/conf.yaml.default

@@ -4,6 +4,33 @@
 
 init_config:
 
+    ## @param file_system_global_exclude - list of strings - optional 
+    ## Instruct the check to always add these patterns to `file_system_exclude`.
+    ##
+    ## WARNING: Overriding these defaults could negatively impact your system or
+    ## the performance of the check.
+    #
+    # file_system_global_exclude:
+    #   - iso9660$
+    #   - tracefs$
+
+    ## @param device_global_exclude - list of strings - optional
+    ## Instruct the check to always add these patterns to `device_exclude`.
+    ##
+    ## WARNING: Overriding these defaults could negatively impact your system or
+    ## the performance of the check.
+    #
+    # device_global_exclude: []
+
+    ## @param mount_point_global_exclude - list of strings - optional
+    ## Instruct the check to always add these patterns to `mount_point_exclude`.
+    ##
+    ## WARNING: Overriding these defaults could negatively impact your system or
+    ## the performance of the check.
+    #
+    # mount_point_global_exclude:
+    #   - (/host)?/proc/sys/fs/binfmt_misc$
+
 instances:
 
     ## @param use_mount - boolean - required
@@ -54,6 +81,89 @@ instances:
     #   - <KEY_1>:<VALUE_1>
     #   - <KEY_2>:<VALUE_2>
 
+    ## @param file_system_include - list of strings - optional
+    ## Instruct the check to only collect from matching file systems.
+    ##
+    ## Character casing is ignored. For convenience, the regular expressions
+    ## start matching from the beginning and therefore to match anywhere you
+    ## must prepend `.*`. For exact matches append `$`.
+    #
+    # file_system_include:
+    #   - ext[34]$
+    #   - ntfs$
+
+    ## @param file_system_exclude - list of strings - optional
+    ## Instruct the check to not collect from matching file systems.
+    ##
+    ## Character casing is ignored. For convenience, the regular expressions
+    ## start matching from the beginning and therefore to match anywhere you
+    ## must prepend `.*`. For exact matches append `$`.
+    ##
+    ## Devices from pseudo or memory-based file systems can be excluded by disabling the
+    ## `include_all_devices` option.
+    ##
+    ## When conflicts arise, this will override `file_system_include`.
+    #
+    # file_system_exclude:
+    #   - tmpfs$
+    #   - rootfs$
+    #   - autofs$
+
+    ## @param device_include - list of strings - optional
+    ## Instruct the check to only collect from matching devices.
+    ##
+    ## Character casing is ignored on Windows. For convenience, the regular
+    ## expressions start matching from the beginning and therefore to match
+    ## anywhere you must prepend `.*`. For exact matches append `$`.
+    #
+    # device_include:
+    #   - /dev/sda[1-3]
+    #   - 'C:'
+
+    ## @param device_exclude - list of strings - optional
+    ## Instruct the check to not collect from matching devices.
+    ##
+    ## Character casing is ignored on Windows. For convenience, the regular
+    ## expressions start matching from the beginning and therefore to match
+    ## anywhere you must prepend `.*`. For exact matches append `$`.
+    ##
+    ## When conflicts arise, this will override `device_include`.
+    #
+    # device_exclude:
+    #   - /dev/sde
+    #   - '[FJ]:'
+
+    ## @param mount_point_include - list of strings - optional
+    ## Instruct the check to only collect from matching mount points.
+    ##
+    ## Character casing is ignored on Windows. For convenience, the regular
+    ## expressions start matching from the beginning and therefore to match
+    ## anywhere you must prepend `.*`. For exact matches append `$`.
+    #
+    # mount_point_include:
+    #   - /dev/sda[1-3]
+    #   - 'C:'
+
+    ## @param mount_point_exclude - list of strings - optional
+    ## Instruct the check to not collect from matching mount points.
+    ##
+    ## Character casing is ignored on Windows. For convenience, the regular
+    ## expressions start matching from the beginning and therefore to match
+    ## anywhere you must prepend `.*`. For exact matches append `$`.
+    #
+    # mount_point_exclude:
+    #   - /proc/sys/fs/binfmt_misc
+    #   - /dev/sde
+    #   - '[FJ]:'
+
+    ## @param include_all_devices - boolean - optional - default: true
+    ## Instruct the check to collect from all devices, including non-physical devices.
+    ## Set this to false to exclude pseudo, memory, duplicate or inaccessible file systems.
+    ##
+    ## For more fine-grained control, use the inclusion and exclusion options.
+    #
+    # include_all_devices: true
+
     ## @param service_check_rw - boolean - optional
     ## Instruct the check to notify based on partition state.
     ##
@@ -68,7 +178,27 @@ instances:
     #
     # tag_by_filesystem: false
 
-    ## @param device_tag_re - list of regex:tags string - optional
+    ## @param tag_by_label - boolean - optional - default: true
+    ## Instruct the check to tag all the metrics with disk label if there is one.
+    ## Works on Linux only.
+    #
+    # tag_by_label: true
+
+    ## @param blkid_cache_file - string - optional
+    ## Instruct the check to read the labels from the blkid cache file instead of `blkid` executable.
+    ## This parameter is used only if `tag_by_label` is true. It is incompatible with `use_lsblk`.
+    ## Works on Linux only.
+    #
+    # blkid_cache_file: /run/blkid/blkid.tab
+
+    ## @param use_lsblk - boolean - optional - default: false
+    ## Instruct the check to read the labels from the `lsblk` executable instead of `blkid` executable.
+    ## This parameter is used only if `tag_by_label` is true. It is incompatible with `blkid_cache_file`.
+    ## Works on Linux only.
+    #
+    # use_lsblk: false
+
+    ## @param device_tag_re - map of regex to tags - optional
     ## Instruct the check to apply additional tags to matching
     ## devices (or mount points if `use_mount` is true).
     ##
@@ -80,3 +210,34 @@ instances:
     #   /san/.*: device_type:san
     #   /dev/sda3: role:db,disk_size:large
     #   "c:": volume:boot
+
+    ## @param min_disk_size - number - optional - default: 0
+    ## Exclude devices with a total disk size less than a minimum value (in MiB)
+    #
+    # min_disk_size: 0
+
+    ## @param timeout - integer - optional - default: 5
+    ## Timeout of the disk query in seconds
+    #
+    # timeout: 5
+
+    ## @param create_mounts - list of mappings - optional
+    ## On Windows, instruct the check to create one or more network
+    ## mounts, and have the check collect metrics for the mounted devices.
+    ##
+    ## Uses the provided username and password (if provided and necessary)
+    ## to create an SMB or NFS mount. If `type` is not specified, then
+    ## the operating system will choose the best available network filesystem
+    ## based on the other parameters. If `type` is specified, then any type
+    ## other than `nfs` will default to an SMB file share.
+    #
+    # create_mounts:
+    #   - mountpoint: 's:'
+    #     user: auser
+    #     password: somepassword
+    #     host: smbserver
+    #     share: space
+    #   - mountpoint: 'n:'
+    #     host: nfsserver
+    #     share: /mnt/nfs_share
+    #     type: nfs

cmd/cluster-agent/subcommands/start/command.go

@@ -121,6 +121,7 @@ import (
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/cluster/orchestrator"
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/cpu/cpu"
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/disk/disk"
+	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/disk/diskv2"
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/disk/io"
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/filehandles"
 	"github.com/DataDog/datadog-agent/pkg/collector/corechecks/system/memory"
@@ -631,7 +632,11 @@ func registerChecks(wlm workloadmeta.Component, tagger tagger.Component, cfg con
 	corecheckLoader.RegisterCheck(kubernetesapiserver.CheckName, kubernetesapiserver.Factory(tagger))
 	corecheckLoader.RegisterCheck(ksm.CheckName, ksm.Factory())
 	corecheckLoader.RegisterCheck(helm.CheckName, helm.Factory())
-	corecheckLoader.RegisterCheck(disk.CheckName, disk.Factory())
+	if cfg.GetBool("use_diskv2_check") {
+		corecheckLoader.RegisterCheck(disk.CheckName, diskv2.Factory())
+	} else {
+		corecheckLoader.RegisterCheck(disk.CheckName, disk.Factory())
+	}
 	corecheckLoader.RegisterCheck(orchestrator.CheckName, orchestrator.Factory(wlm, cfg, tagger))
 	corecheckLoader.RegisterCheck(winproc.CheckName, winproc.Factory())
 }

pkg/collector/corechecks/system/disk/diskv2/clock.go

@@ -0,0 +1,33 @@
+// Unless explicitly stated otherwise all files in this repository are licensed
+// under the Apache License Version 2.0.
+// This product includes software developed at Datadog (https://www.datadoghq.com/).
+// Copyright 2016-present Datadog, Inc.
+
+// Package diskv2 provides Disk Check.
+package diskv2
+
+import "time"
+
+// Clock abstracts time-related functions used in the package.
+type Clock interface {
+	// After returns a channel that will send the current time after the specified duration.
+	After(d time.Duration) <-chan time.Time
+}
+
+// RealClock is the production implementation that wraps time.After.
+type RealClock struct{}
+
+// After returns the channel from time.After.
+func (rc *RealClock) After(d time.Duration) <-chan time.Time {
+	return time.After(d)
+}
+
+// FakeClock is used in unit tests.
+type FakeClock struct {
+	AfterCh chan time.Time
+}
+
+// After returns the controlled channel.
+func (fc *FakeClock) After(_ time.Duration) <-chan time.Time {
+	return fc.AfterCh
+}