NFS
NFS metrics collector that collects the following data:
- RPC throughput metrics
- NFS mount point metrics (NFSv3 and v4 only)
- NFSd throughput metrics
Configuration¶
Preconditions¶
- The NFS client environment is properly configured.
- The NFS client is properly mounted to the server's shared directory.
Collector Configuration¶
Go to the conf.d/host directory under the DataKit installation directory, copy nfs.conf.sample and name it nfs.conf. Examples are as follows:
[[inputs.nfs]]
##(optional) collect interval, default is 10 seconds
interval = '10s'
## Whether to enable NFSd metric collection
# nfsd = true
## NFS mount point metric configuration
[inputs.nfs.mountstats]
## Enable r/w statistics
# rw = true
## Enable transport statistics
# transport = true
## Enable event statistics
# event = true
## Enable operation statistics
# operations = true
[inputs.nfs.tags]
# some_tag = "some_value"
# more_tag = "some_other_value"
Once configured, restart DataKit.
Can be turned on by ConfigMap Injection Collector Configuration or Config ENV_DATAKIT_INPUTS .
Can also be turned on by environment variables, (needs to be added as the default collector in ENV_DEFAULT_ENABLED_INPUTS):
-
ENV_INPUT_NFS_INTERVAL
采集器重复间隔时长
字段类型: Duration
采集器配置字段:
interval默认值:
10s -
ENV_INPUT_NFS_ENABLE_MOUNT_STATS_RW_BYTES
开启 NFS 挂载点的详细字节读写信息
字段类型: Boolean
采集器配置字段:
enable_mount_stats_rw_bytes默认值:
false -
ENV_INPUT_NFS_ENABLE_MOUNT_STATS_TRANSPORT
开启 NFS 挂载点与服务端的传输信息
字段类型: Boolean
采集器配置字段:
enable_mount_stats_transport默认值:
false -
ENV_INPUT_NFS_ENABLE_MOUNT_STATS_EVENT
开启 NFS 事件统计信息
字段类型: Boolean
采集器配置字段:
enable_mount_stats_event默认值:
false -
ENV_INPUT_NFS_ENABLE_MOUNT_STATS_OPERATIONS
开启 NFS 给定操作的传输信息
字段类型: Boolean
采集器配置字段:
enable_mount_stats_operations默认值:
false -
ENV_INPUT_NFS_NFSD
开启 NFSd 指标
字段类型: Boolean
采集器配置字段:
nfsd默认值:
false
NFSd Start¶
NFSd is the daemon of the NFS service, a key component on the server side, responsible for handling NFS requests sent by clients. If the local machine is also used as an NFS server, you can enable this metric to view statistics such as network, disk I/O, and threads on which users process NFS requests.
If you want to enable it, you need to modify the configuration file.
[[inputs.nfs]]
##(optional) collect interval, default is 10 seconds
interval = '10s'
## Whether to enable NFSd metric collection
nfsd = true
...
NFS mount point stats Start¶
The default set of nfs_mountstats metrics only displays statistics about the disk usage and NFS running time of the mount point, and you need to modify the configuration file to view the R/W, Transport, Event, and Operations information of the NFS mount point.
[[inputs.nfs]]
...
## NFS mount point metric configuration
[inputs.nfs.mountstats]
## Enable r/w statistics
rw = true
## Enable transport statistics
transport = true
## Enable event statistics
event = true
## Enable operation statistics
operations = true
...
Metric¶
For all of the following data collections, a global tag named host is appended by default (the tag value is the host name of the DataKit), or other tags can be specified in the configuration by [inputs.nfs.tags]:
nfs¶
- Tags
| Tag | Description |
|---|---|
| method | Invoked method. |
| protocol | Protocol type. |
- Metrics
| Metric | Description |
|---|---|
| connections_total | Total number of NFSd TCP connections. Type: int Unit: count |
| requests_total | Number of NFS procedures invoked. Type: int Unit: count |
| rpc_authentication_refreshes_total | Number of RPC authentication refreshes performed. Type: int Unit: count |
| rpc_retransmissions_total | Number of RPC transmissions performed. Type: int Unit: count |
| rpcs_total | Total number of RPCs performed. Type: int Unit: count |
| tcp_packets_total | Total NFSd network TCP packets (sent+received) by protocol type. Type: int Unit: count |
| udp_packets_total | Total NFSd network UDP packets (sent+received) by protocol type. Type: int Unit: count |
nfs_mountstats¶
- Tags
| Tag | Description |
|---|---|
| device | Device name. |
| mountpoint | Where the device is mounted. |
| type | Device type. |
- Metrics
| Metric | Description |
|---|---|
| age_seconds_total | The age of the NFS mount in seconds. Type: int Unit: time,s |
| direct_read_bytes_total | Number of bytes read using the read() syscall in O_DIRECT mode. Type: int Unit: digital,B |
| direct_write_bytes_total | Number of bytes written using the write() syscall in O_DIRECT mode. Type: int Unit: digital,B |
| event_attribute_invalidate_total | Number of times cached inode attributes are invalidated. Type: int Unit: count |
| event_data_invalidate_total | Number of times an inode cache is cleared. Type: int Unit: count |
| event_dnode_revalidate_total | Number of times cached dentry nodes are re-validated from the server. Type: int Unit: count |
| event_inode_revalidate_total | Number of times cached inode attributes are re-validated from the server. Type: int Unit: count |
| event_jukebox_delay_total | Number of times the NFS server indicated EJUKEBOX; retrieving data from offline storage. Type: int Unit: count |
| event_pnfs_read_total | Number of NFS v4.1+ pNFS reads. Type: int Unit: count |
| event_pnfs_write_total | Number of NFS v4.1+ pNFS writes. Type: int Unit: count |
| event_short_read_total | Number of times the NFS server gave less data than expected while reading. Type: int Unit: count |
| event_short_write_total | Number of times the NFS server wrote less data than expected while writing. Type: int Unit: count |
| event_silly_rename_total | Number of times a file was removed while still open by another process. Type: int Unit: count |
| event_truncation_total | Number of times files have been truncated. Type: int Unit: count |
| event_vfs_access_total | Number of times permissions have been checked. Type: int Unit: count |
| event_vfs_file_release_total | Number of times files have been closed and released. Type: int Unit: count |
| event_vfs_flush_total | Number of pending writes that have been forcefully flushed to the server. Type: int Unit: count |
| event_vfs_fsync_total | Number of times fsync() has been called on directories and files. Type: int Unit: count |
| event_vfs_getdents_total | Number of times directory entries have been read with getdents(). Type: int Unit: count |
| event_vfs_lock_total | Number of times locking has been attempted on a file. Type: int Unit: count |
| event_vfs_lookup_total | Number of times a directory lookup has occurred. Type: int Unit: count |
| event_vfs_open_total | Number of times cached inode attributes are invalidated. Type: int Unit: count |
| event_vfs_read_page_total | Number of pages read directly via mmap()'d files. Type: int Unit: count |
| event_vfs_read_pages_total | Number of times a group of pages have been read. Type: int Unit: count |
| event_vfs_setattr_total | Number of times directory entries have been read with getdents(). Type: int Unit: count |
| event_vfs_update_page_total | Number of updates (and potential writes) to pages. Type: int Unit: count |
| event_vfs_write_page_total | Number of pages written directly via mmap()'d files. Type: int Unit: count |
| event_vfs_write_pages_total | Number of times a group of pages have been written. Type: int Unit: count |
| event_write_extension_total | Number of times a file has been grown due to writes beyond its existing end. Type: int Unit: count |
| fs_avail | Available space on the filesystem. Type: int Unit: digital,B |
| fs_size | Total size of the filesystem. Type: int Unit: digital,B |
| fs_used | Used space on the filesystem. Type: int Unit: digital,B |
| fs_used_percent | Percentage of used space on the filesystem. Type: float Unit: percent,percent |
| operations_latency_seconds | Average RPC latency (RTT) for a given operation, in seconds. Type: float Unit: time,s |
| operations_major_timeouts_total | Number of times a request has had a major timeout for a given operation. Type: int Unit: count |
| operations_queue_time_seconds_total | Duration all requests spent queued for transmission for a given operation before they were sent, in seconds. Type: time,s Unit: count |
| operations_received_bytes_total | Number of bytes received for a given operation, including RPC headers and payload. Type: int Unit: digital,B |
| operations_request_time_seconds_total | Duration all requests took from when a request was enqueued to when it was completely handled for a given operation, in seconds. Type: time,s Unit: count |
| operations_requests_total | Number of requests performed for a given operation. Type: int Unit: count |
| operations_response_time_seconds_total | Duration all requests took to get a reply back after a request for a given operation was transmitted, in seconds. Type: time,s Unit: count |
| operations_sent_bytes_total | Number of bytes sent for a given operation, including RPC headers and payload. Type: int Unit: digital,B |
| operations_transmissions_total | Number of times an actual RPC request has been transmitted for a given operation. Type: int Unit: count |
| read_bytes_total | Number of bytes read using the read() syscall. Type: int Unit: digital,B |
| read_pages_total | Number of pages read directly via mmap()'d files. Type: int Unit: count |
| total_read_bytes_total | Number of bytes read from the NFS server, in total. Type: int Unit: digital,B |
| total_write_bytes_total | Number of bytes written to the NFS server, in total. Type: int Unit: digital,B |
| transport_backlog_queue_total | Total number of items added to the RPC backlog queue. Type: int Unit: count |
| transport_bad_transaction_ids_total | Number of times the NFS server sent a response with a transaction ID unknown to this client. Type: int Unit: count |
| transport_bind_total | Number of times the client has had to establish a connection from scratch to the NFS server. Type: int Unit: count |
| transport_connect_total | Number of times the client has made a TCP connection to the NFS server. Type: int Unit: count |
| transport_idle_time_seconds | Duration since the NFS mount last saw any RPC traffic, in seconds. Type: int Unit: time,s |
| transport_maximum_rpc_slots | Maximum number of simultaneously active RPC requests ever used. Type: int Unit: count |
| transport_pending_queue_total | Total number of items added to the RPC transmission pending queue. Type: int Unit: count |
| transport_receives_total | Number of RPC responses for this mount received from the NFS server. Type: int Unit: count |
| transport_sending_queue_total | Total number of items added to the RPC transmission sending queue. Type: int Unit: count |
| transport_sends_total | Number of RPC requests for this mount sent to the NFS server. Type: int Unit: count |
| write_bytes_total | Number of bytes written using the write() syscall. Type: int Unit: digital,B |
| write_pages_total | Number of pages written directly via mmap()'d files. Type: int Unit: count |
nfsd¶
- Tags
| Tag | Description |
|---|---|
| error | Error type. |
| method | Invoked method. |
| protocol | Protocol type. |
- Metrics
| Metric | Description |
|---|---|
| connections_total | Total number of NFSd TCP connections. Type: int Unit: count |
| disk_bytes_read_total | Total NFSd bytes read. Type: int Unit: count |
| disk_bytes_written_total | Total NFSd bytes written. Type: int Unit: count |
| file_handles_stale_total | Total number of NFSd stale file handles Type: int Unit: count |
| read_ahead_cache_not_found_total | Total number of NFSd read ahead cache not found. Type: int Unit: count |
| read_ahead_cache_size_blocks | How large the read ahead cache is in blocks. Type: int Unit: count |
| reply_cache_hits_total | Total number of NFSd Reply Cache hits (client lost server response). Type: int Unit: count |
| reply_cache_misses_total | Total number of NFSd Reply Cache an operation that requires caching (idempotent). Type: int Unit: count |
| reply_cache_nocache_total | Total number of NFSd Reply Cache non-idempotent operations (rename/delete/…). Type: int Unit: count |
| rpc_errors_total | Total number of NFSd RPC errors by error type. Type: int Unit: count |
| server_rpcs_total | Total number of NFSd RPCs. Type: int Unit: count |
| server_threads | Total number of NFSd kernel threads that are running. Type: int Unit: count |
| tcp_packets_total | Total NFSd network TCP packets (sent+received) by protocol type. Type: int Unit: count |
| udp_packets_total | Total NFSd network UDP packets (sent+received) by protocol type. Type: int Unit: count |