smart

S.M.A.R.T. 插件

从telegraf fork，略作改动

Get metrics using the command line utility smartctl for S.M.A.R.T. (Self-Monitoring, Analysis and Reporting Technology) storage devices. SMART is a monitoring system included in computer hard disk drives (HDDs) and solid-state drives (SSDs) that detects and reports on various indicators of drive reliability, with the intent of enabling the anticipation of hardware failures. See smartmontools (https://www.smartmontools.org/).

SMART information is separated between different measurements: smart_device is used for general information, while smart_attribute stores the detailed attribute information if attributes = true is enabled in the plugin configuration.

If no devices are specified, the plugin will scan for SMART devices via the following command:

smartctl --scan

Metrics will be reported from the following smartctl command:

smartctl --info --attributes --health -n <nocheck> --format=brief <device>

This plugin supports smartmontools version 5.41 and above, but v. 5.41 and v. 5.42 might require setting nocheck, see the comment in the sample configuration. Also, NVMe capabilities were introduced in version 6.5.

To enable SMART on a storage device run:

smartctl -s on <device>

NVMe vendor specific attributes

For NVMe disk type, plugin can use command line utility nvme-cli. It has a feature to easy access a vendor specific attributes. This plugin supports nmve-cli version 1.5 and above (https://github.com/linux-nvme/nvme-cli). In case of nvme-cli absence NVMe vendor specific metrics will not be obtained.

Vendor specific SMART metrics for NVMe disks may be reported from the following nvme command:

nvme <vendor> smart-log-add <device>

Note that vendor plugins for nvme-cli could require different naming convention and report format.

To see installed plugin extensions, depended on the nvme-cli version, look at the bottom of:

nvme help

To gather disk vendor id (vid) id-ctrl could be used:

nvme id-ctrl <device>

Association between a vid and company can be found there: https://pcisig.com/membership/member-companies.

Devices affiliation to being NVMe or non NVMe will be determined thanks to:

smartctl --scan

and:

smartctl --scan -d nvme

Configuration

# Read metrics from storage devices supporting S.M.A.R.T.
[[instances]]
## Optionally specify the path to the smartctl executable
# path_smartctl = "/usr/bin/smartctl"

## Optionally specify the path to the nvme-cli executable
# path_nvme = "/usr/bin/nvme"

## Optionally specify if vendor specific attributes should be propagated for NVMe disk case
## ["auto-on"] - automatically find and enable additional vendor specific disk info
## ["vendor1", "vendor2", ...] - e.g. "Intel" enable additional Intel specific disk info
# enable_extensions = ["auto-on"]

## On most platforms used cli utilities requires root access.
## Setting 'use_sudo' to true will make use of sudo to run smartctl or nvme-cli.
## Sudo must be configured to allow the categraf user to run smartctl or nvme-cli
## Sudo must be configured to allow the categraf user to run smartctl or nvme-cli
## without a password.
use_sudo = true

## Skip checking disks in this power mode. Defaults to
## "standby" to not wake up disks that have stopped rotating.
## See --nocheck in the man pages for smartctl.
## smartctl version 5.41 and 5.42 have faulty detection of
## power mode and might require changing this value to
## "never" depending on your disks.
# nocheck = "standby"

## Gather all returned S.M.A.R.T. attribute metrics and the detailed
## information from each drive into the 'smart_attribute' measurement.
attributes = true

## Optionally specify devices to exclude from reporting if disks auto-discovery is performed.
# excludes = [ "/dev/pass6" ]

## Optionally specify devices and device type, if unset
## a scan (smartctl --scan and smartctl --scan -d nvme) for S.M.A.R.T. devices will be done
## and all found will be included except for the excluded in excludes.
# devices = [ "/dev/ada0 -d atacam", "/dev/nvme0"]
# devices = ["dev/nvme0 -d nvme", "/dev/nvme0"]

## Timeout for the cli command to complete.
timeout = "30s"

## Optionally call smartctl and nvme-cli with a specific concurrency policy.
## By default, smartctl and nvme-cli are called in separate threads (goroutines) to gather disk attributes.
## Some devices (e.g. disks in RAID arrays) may have access limitations that require sequential reading of
## SMART data - one individual array drive at the time. In such case please set this configuration option
## to "sequential" to get readings for all drives.
## valid options: concurrent, sequential
# read_method = "concurrent"

Permissions

采集需要sudo权限

Metrics

• smart_device:

  • tags:
    • capacity
    • device
    • enabled
    • model
    • serial_no
    • wwn
  • fields:
    • exit_status
    • health_ok
    • media_wearout_indicator
    • percent_lifetime_remain
    • read_error_rate
    • seek_error
    • temp_c
    • udma_crc_errors
    • wear_leveling_count

• smart_attribute:

  • tags:
    • capacity
    • device
    • enabled
    • fail
    • flags
    • id
    • model
    • name
    • serial_no
    • wwn
  • fields:
    • exit_status
    • threshold
    • value
    • worst
    • critical_warning
    • temperature_celsius
    • available_spare
    • available_spare_threshold
    • percentage_used
    • data_units_read
    • data_units_written
    • host_read_commands
    • host_write_commands
    • controller_busy_time
    • power_cycle_count
    • power_on_hours
    • unsafe_shutdowns
    • media_and_data_integrity_errors
    • error_information_log_entries
    • warning_temperature_time
    • critical_temperature_time
    • program_fail_count
    • erase_fail_count
    • wear_leveling_count
    • end_to_end_error_detection_count
    • crc_error_count
    • media_wear_percentage
    • host_reads
    • timed_workload_timer
    • thermal_throttle_status
    • retry_buffer_overflow_count
    • pll_lock_loss_count

Flags

The interpretation of the tag flags is:

• K auto-keep
• C event count
• R error rate
• S speed/performance
• O updated online
• P prefailure warning

Exit Status

The exit_status field captures the exit status of the used cli utilities command which is defined by a bitmask. For the interpretation of the bitmask see the man page for smartctl or nvme-cli.

Device Names

Device names, e.g., /dev/sda, are not persistent, and may be subject to change across reboots or system changes. Instead, you can use the World Wide Name (WWN) or serial number to identify devices. On Linux block devices can be referenced by the WWN in the following location: /dev/disk/by-id/.

Troubleshooting

If you expect to see more SMART metrics than this plugin shows, be sure to use a proper version of smartctl or nvme-cli utility which has the functionality to gather desired data. Also, check your device capability because not every SMART metrics are mandatory. For example the number of temperature sensors depends on the device specification.

If this plugin is not working as expected for your SMART enabled device, please run these commands and include the output in a bug report:

For non NVMe devices (from smartctl version >= 7.0 this will also return NVMe devices by default):

smartctl --scan

For NVMe devices:

smartctl --scan -d nvme

Run the following command replacing your configuration setting for NOCHECK and the DEVICE (name of the device could be taken from the previous command):

smartctl --info --health --attributes --tolerance=verypermissive --nocheck NOCHECK --format=brief -d DEVICE

If you try to gather vendor specific metrics, please provide this command and replace vendor and device to match your case:

nvme VENDOR smart-log-add DEVICE

If you have specified devices array in configuration file, and categraf only shows data from one device, you should change the plugin configuration to sequentially gather disk attributes instead of collecting it in separate threads (goroutines). To do this find in plugin configuration read_method and change it to sequential:

    ## Optionally call smartctl and nvme-cli with a specific concurrency policy.
    ## By default, smartctl and nvme-cli are called in separate threads (goroutines) to gather disk attributes.
    ## Some devices (e.g. disks in RAID arrays) may have access limitations that require sequential reading of
    ## SMART data - one individual array drive at the time. In such case please set this configuration option
    ## to "sequential" to get readings for all drives.
    ## valid options: concurrent, sequential
    read_method = "sequential"

Example Output

smart_device_health_ok agent_hostname=1.2.3.4 device=nvme0 model=INTEL_SSDPE2KX040T8 serial_no=PHLJ830200CH4P0DGN 1
smart_device_temp_c agent_hostname=1.2.3.4 device=nvme0 model=INTEL_SSDPE2KX040T8 serial_no=PHLJ830200CH4P0DGN 53
smart_attribute_program_fail_count agent_hostname=1.2.3.4 device=nvme0 model= name=Program_Fail_Count serial_no=PHLJ830200CH4P0DGN 0
smart_attribute_erase_fail_count agent_hostname=1.2.3.4 device=nvme0 model= name=Erase_Fail_Count serial_no=PHLJ830200CH4P0DGN 0
smart_attribute_wear_leveling_count agent_hostname=1.2.3.4 device=nvme0 model= name=Wear_Leveling_Count serial_no=PHLJ830200CH4P0DGN 34360328200