17 KiB

Raw Blame History

Node exporter

Prometheus exporter for hardware and OS metrics exposed by *NIX kernels, written in Go with pluggable metric collectors.

The Windows exporter is recommended for Windows users. To expose NVIDIA GPU metrics, prometheus-dcgm can be used.

Installation and Usage

If you are new to Prometheus and node_exporter there is a simple step-by-step guide.

The node_exporter listens on HTTP port 9100 by default. See the --help output for more options.

Ansible

For automated installs with Ansible, there is the Cloud Alchemy role.

RHEL/CentOS/Fedora

There is a community-supplied COPR repository which closely follows upstream releases.

Docker

The node_exporter is designed to monitor the host system. It's not recommended to deploy it as a Docker container because it requires access to the host system.

For situations where Docker deployment is needed, some extra flags must be used to allow the node_exporter access to the host namespaces.

Be aware that any non-root mount points you want to monitor will need to be bind-mounted into the container.

If you start container for host monitoring, specify path.rootfs argument. This argument must match path in bind-mount of host root. The node_exporter will use path.rootfs as prefix to access host filesystem.

docker run -d \
  --net="host" \
  --pid="host" \
  -v "/:/host:ro,rslave" \
  quay.io/prometheus/node-exporter:latest \
  --path.rootfs=/host

For Docker compose, similar flag changes are needed.

---
version: '3.8'

services:
  node_exporter:
    image: quay.io/prometheus/node-exporter:latest
    container_name: node_exporter
    command:
      - '--path.rootfs=/host'
    network_mode: host
    pid: host
    restart: unless-stopped
    volumes:
      - '/:/host:ro,rslave'

On some systems, the timex collector requires an additional Docker flag, --cap-add=SYS_TIME, in order to access the required syscalls.

Collectors

There is varying support for collectors on each operating system. The tables below list all existing collectors and the supported systems.

Collectors are enabled by providing a --collector.<name> flag. Collectors that are enabled by default can be disabled by providing a --no-collector.<name> flag. To enable only some specific collector(s), use --collector.disable-defaults --collector.<name> ....

Enabled by default

Name	Description	OS
arp	Exposes ARP statistics from `/proc/net/arp`.	Linux
bcache	Exposes bcache statistics from `/sys/fs/bcache/`.	Linux
bonding	Exposes the number of configured and active slaves of Linux bonding interfaces.	Linux
btrfs	Exposes btrfs statistics	Linux
boottime	Exposes system boot time derived from the `kern.boottime` sysctl.	Darwin, Dragonfly, FreeBSD, NetBSD, OpenBSD, Solaris
conntrack	Shows conntrack statistics (does nothing if no `/proc/sys/net/netfilter/` present).	Linux
cpu	Exposes CPU statistics	Darwin, Dragonfly, FreeBSD, Linux, Solaris, OpenBSD
cpufreq	Exposes CPU frequency statistics	Linux, Solaris
diskstats	Exposes disk I/O statistics.	Darwin, Linux, OpenBSD
dmi	Expose Desktop Management Interface (DMI) info from `/sys/class/dmi/id/`	Linux
edac	Exposes error detection and correction statistics.	Linux
entropy	Exposes available entropy.	Linux
exec	Exposes execution statistics.	Dragonfly, FreeBSD
fibrechannel	Exposes fibre channel information and statistics from `/sys/class/fc_host/`.	Linux
filefd	Exposes file descriptor statistics from `/proc/sys/fs/file-nr`.	Linux
filesystem	Exposes filesystem statistics, such as disk space used.	Darwin, Dragonfly, FreeBSD, Linux, OpenBSD
hwmon	Expose hardware monitoring and sensor data from `/sys/class/hwmon/`.	Linux
infiniband	Exposes network statistics specific to InfiniBand and Intel OmniPath configurations.	Linux
ipvs	Exposes IPVS status from `/proc/net/ip_vs` and stats from `/proc/net/ip_vs_stats`.	Linux
loadavg	Exposes load average.	Darwin, Dragonfly, FreeBSD, Linux, NetBSD, OpenBSD, Solaris
mdadm	Exposes statistics about devices in `/proc/mdstat` (does nothing if no `/proc/mdstat` present).	Linux
meminfo	Exposes memory statistics.	Darwin, Dragonfly, FreeBSD, Linux, OpenBSD
netclass	Exposes network interface info from `/sys/class/net/`	Linux
netdev	Exposes network interface statistics such as bytes transferred.	Darwin, Dragonfly, FreeBSD, Linux, OpenBSD
netstat	Exposes network statistics from `/proc/net/netstat`. This is the same information as `netstat -s`.	Linux
nfs	Exposes NFS client statistics from `/proc/net/rpc/nfs`. This is the same information as `nfsstat -c`.	Linux
nfsd	Exposes NFS kernel server statistics from `/proc/net/rpc/nfsd`. This is the same information as `nfsstat -s`.	Linux
nvme	Exposes NVMe info from `/sys/class/nvme/`	Linux
os	Expose OS release info from `/etc/os-release` or `/usr/lib/os-release`	any
powersupplyclass	Exposes Power Supply statistics from `/sys/class/power_supply`	Linux
pressure	Exposes pressure stall statistics from `/proc/pressure/`.	Linux (kernel 4.20+ and/or CONFIG_PSI)
rapl	Exposes various statistics from `/sys/class/powercap`.	Linux
schedstat	Exposes task scheduler statistics from `/proc/schedstat`.	Linux
selinux	Exposes SELinux statistics.	Linux
sockstat	Exposes various statistics from `/proc/net/sockstat`.	Linux
softnet	Exposes statistics from `/proc/net/softnet_stat`.	Linux
stat	Exposes various statistics from `/proc/stat`. This includes boot time, forks and interrupts.	Linux
tapestats	Exposes statistics from `/sys/class/scsi_tape`.	Linux
textfile	Exposes statistics read from local disk. The `--collector.textfile.directory` flag must be set.	any
thermal	Exposes thermal statistics like `pmset -g therm`.	Darwin
thermal_zone	Exposes thermal zone & cooling device statistics from `/sys/class/thermal`.	Linux
time	Exposes the current system time.	any
timex	Exposes selected adjtimex(2) system call stats.	Linux
udp_queues	Exposes UDP total lengths of the rx_queue and tx_queue from `/proc/net/udp` and `/proc/net/udp6`.	Linux
uname	Exposes system information as provided by the uname system call.	Darwin, FreeBSD, Linux, OpenBSD
vmstat	Exposes statistics from `/proc/vmstat`.	Linux
xfs	Exposes XFS runtime statistics.	Linux (kernel 4.4+)
zfs	Exposes ZFS performance statistics.	FreeBSD, Linux, Solaris

Disabled by default

node_exporter also implements a number of collectors that are disabled by default. Reasons for this vary by collector, and may include:

High cardinality
Prolonged runtime that exceeds the Prometheus scrape_interval or scrape_timeout
Significant resource demands on the host

You can enable additional collectors as desired by adding them to your init system's or service supervisor's startup configuration for node_exporter but caution is advised. Enable at most one at a time, testing first on a non-production system, then by hand on a single production node. When enabling additional collectors, you should carefully monitor the change by observing the scrape_duration_seconds metric to ensure that collection completes and does not time out. In addition, monitor the scrape_samples_post_metric_relabeling metric to see the changes in cardinality.

Name	Description	OS
buddyinfo	Exposes statistics of memory fragments as reported by /proc/buddyinfo.	Linux
cgroups	A summary of the number of active and enabled cgroups	Linux
devstat	Exposes device statistics	Dragonfly, FreeBSD
drbd	Exposes Distributed Replicated Block Device statistics (to version 8.4)	Linux
ethtool	Exposes network interface information and network driver statistics equivalent to `ethtool`, `ethtool -S`, and `ethtool -i`.	Linux
interrupts	Exposes detailed interrupts statistics.	Linux, OpenBSD
ksmd	Exposes kernel and system statistics from `/sys/kernel/mm/ksm`.	Linux
lnstat	Exposes stats from `/proc/net/stat/`.	Linux
logind	Exposes session counts from logind.	Linux
meminfo_numa	Exposes memory statistics from `/proc/meminfo_numa`.	Linux
mountstats	Exposes filesystem statistics from `/proc/self/mountstats`. Exposes detailed NFS client statistics.	Linux
network_route	Exposes the routing table as metrics	Linux
ntp	Exposes local NTP daemon health to check time	any
perf	Exposes perf based metrics (Warning: Metrics are dependent on kernel configuration and settings).	Linux
processes	Exposes aggregate process statistics from `/proc`.	Linux
qdisc	Exposes queuing discipline statistics	Linux
runit	Exposes service status from runit.	any
slabinfo	Exposes slab statistics from `/proc/slabinfo`. Note that permission of `/proc/slabinfo` is usually 0400, so set it appropriately.	Linux
supervisord	Exposes service status from supervisord.	any
sysctl	Expose sysctl values from `/proc/sys`. Use `--collector.sysctl.include(-info)` to configure.	Linux
systemd	Exposes service and system status from systemd.	Linux
tcpstat	Exposes TCP connection status information from `/proc/net/tcp` and `/proc/net/tcp6`. (Warning: the current version has potential performance issues in high load situations.)	Linux
wifi	Exposes WiFi device and station statistics.	Linux
zoneinfo	Exposes NUMA memory zone metrics.	Linux

Perf Collector

The perf collector may not work out of the box on some Linux systems due to kernel configuration and security settings. To allow access, set the following sysctl parameter:

sysctl -w kernel.perf_event_paranoid=X

2 allow only user-space measurements (default since Linux 4.6).
1 allow both kernel and user measurements (default before Linux 4.6).
0 allow access to CPU-specific data but not raw tracepoint samples.
-1 no restrictions.

Depending on the configured value different metrics will be available, for most cases 0 will provide the most complete set. For more information see man 2 perf_event_open.

By default, the perf collector will only collect metrics of the CPUs that node_exporter is running on (ie runtime.NumCPU. If this is insufficient (e.g. if you run node_exporter with its CPU affinity set to specific CPUs), you can specify a list of alternate CPUs by using the --collector.perf.cpus flag. For example, to collect metrics on CPUs 2-6, you would specify: --collector.perf --collector.perf.cpus=2-6. The CPU configuration is zero indexed and can also take a stride value; e.g. --collector.perf --collector.perf.cpus=1-10:5 would collect on CPUs 1, 5, and 10.

The perf collector is also able to collect tracepoint counts when using the --collector.perf.tracepoint flag. Tracepoints can be found using perf list or from debugfs. And example usage of this would be --collector.perf.tracepoint="sched:sched_process_exec".

Sysctl Collector

The sysctl collector can be enabled with --collector.sysctl. It supports exposing numeric sysctl values as metrics using the --collector.sysctl.include flag and string values as info metrics by using the --collector.sysctl.include-info flag. The flags can be repeated. For sysctl with multiple numeric values, an optional mapping can be given to expose each value as its own metric. Otherwise an index label is used to identify the different fields.

Examples

Numeric values

Single values

Using --collector.sysctl.include=vm.user_reserve_kbytes: vm.user_reserve_kbytes = 131072 -> node_sysctl_vm_user_reserve_kbytes 131072

Multiple values

A sysctl can contain multiple values, for example:

net.ipv4.tcp_rmem = 4096	131072	6291456

Using --collector.sysctl.include=net.ipv4.tcp_rmem the collector will expose:

node_sysctl_net_ipv4_tcp_rmem{index="0"} 4096
node_sysctl_net_ipv4_tcp_rmem{index="1"} 131072
node_sysctl_net_ipv4_tcp_rmem{index="2"} 6291456

If the indexes have defined meaning like in this case, the values can be mapped to multiple metrics by appending the mapping to the --collector.sysctl.include flag: Using --collector.sysctl.include=net.ipv4.tcp_rmem:min,default,max the collector will expose:

node_sysctl_net_ipv4_tcp_rmem_min 4096
node_sysctl_net_ipv4_tcp_rmem_default 131072
node_sysctl_net_ipv4_tcp_rmem_max 6291456

String values

String values need to be exposed as info metric. The user selects them by using the --collector.sysctl.include-info flag.

Single values

kernel.core_pattern = core -> node_sysctl_info{key="kernel.core_pattern_info", value="core"} 1

Multiple values

Given the following sysctl:

kernel.seccomp.actions_avail = kill_process kill_thread trap errno trace log allow

Setting --collector.sysctl.include-info=kernel.seccomp.actions_avail will yield:

node_sysctl_info{key="kernel.seccomp.actions_avail", index="0", value="kill_process"} 1
node_sysctl_info{key="kernel.seccomp.actions_avail", index="1", value="kill_thread"} 1
...

Textfile Collector

The textfile collector is similar to the Pushgateway, in that it allows exporting of statistics from batch jobs. It can also be used to export static metrics, such as what role a machine has. The Pushgateway should be used for service-level metrics. The textfile module is for metrics that are tied to a machine.

To use it, set the --collector.textfile.directory flag on the node_exporter commandline. The collector will parse all files in that directory matching the glob *.prom using the text format. Note: Timestamps are not supported.

To atomically push completion time for a cron job:

echo my_batch_job_completion_time $(date +%s) > /path/to/directory/my_batch_job.prom.$$
mv /path/to/directory/my_batch_job.prom.$$ /path/to/directory/my_batch_job.prom

To statically set roles for a machine using labels:

echo 'role{role="application_server"} 1' > /path/to/directory/role.prom.$$
mv /path/to/directory/role.prom.$$ /path/to/directory/role.prom

Filtering enabled collectors

The node_exporter will expose all metrics from enabled collectors by default. This is the recommended way to collect metrics to avoid errors when comparing metrics of different families.

For advanced use the node_exporter can be passed an optional list of collectors to filter metrics. The collect[] parameter may be used multiple times. In Prometheus configuration you can use this syntax under the scrape config.

  params:
    collect[]:
      - foo
      - bar

This can be useful for having different Prometheus servers collect specific metrics from nodes.

Development building and running

Prerequisites:

Go compiler
RHEL/CentOS: glibc-static package.

Building:

git clone https://github.com/prometheus/node_exporter.git
cd node_exporter
make build
./node_exporter <flags>

To see all available configuration flags:

./node_exporter -h

Running tests

make test

TLS endpoint

** EXPERIMENTAL **

The exporter supports TLS via a new web configuration file.

./node_exporter --web.config=web-config.yml

See the exporter-toolkit https package for more details.

17 KiB Raw Blame History

Node exporter

Installation and Usage

Ansible

RHEL/CentOS/Fedora

Docker

Collectors

Enabled by default

Disabled by default

Perf Collector

Sysctl Collector

Examples

Numeric values

Single values

Multiple values

String values

Single values

Multiple values

Textfile Collector

Filtering enabled collectors

Development building and running

Running tests

TLS endpoint

17 KiB

Raw Blame History