github prometheus/node_exporter v1.13.0
1.13.0 / 2026-11-19

one hour ago

NOTE:

The behavior of node_cpu_frequency_hertz has changed. There are two different sources
of data for this metric. One from the cpu collector, one from the cpufreq collector.
Due to historical feature migrations the cpufreq collector shares the same subsystem
as the cpu collector. This caused conflicting data behavior. An attempt to fix this
was introduced in #3318. This behavior has been reverted to the previous behavior in #3795.

A new flag (--collector.cpufreq.enable-cpufreq-prefix) has been introduced that gives
the cpufreq a dedicated metric name subsystem (node_cpufreq_...). This will likely
be the default in v2.

  • [CHANGE] netstat: Use procfs parsers for netstat, snmp and snmp6 #3796
  • [CHANGE] edac: Fix no data results #3863
  • [CHANGE] nvme: Deprecate metric nvme_namespace_capacity_bytes #3899
  • [FEATURE] Add tainted collector for /proc/sys/kernel/tainted #3745
  • [FEATURE] Add resctrl collector #3797
  • [ENHANCEMENT] netstat: Add new FreeBSD metrics #3334
  • [ENHANCEMENT] Add chip label to DRM metrics #3572
  • [ENHANCEMENT] diskstats: add device-mapper backing device info metric #3676
  • [ENHANCEMENT] textfile: support UTF-8 metric and label names #3729
  • [ENHANCEMENT] ksmd: add general_profit metric #3778 #3894
  • [BUGFIX] edac: Fix printf-style formatting in slog call #3759
  • [BUGFIX] fibre_channel: fix crash when attempting to dereference invalid counters #3573
  • [BUGFIX] thermal: Fix collecting Darwin temperatures without CPU power status #3767
  • [BUGFIX] fibrechannel: Fix for ppc64le #3769
  • [BUGFIX] cpu: Revert dup cpu frequency, add node_cpufreq flag #3795
  • [BUGFIX] infiniband: skip sysfs reads for excluded devices #3824
  • [BUGFIX] bcache: Fix btree_read_average_duration_seconds unit #3828

Don't miss a new node_exporter release

NewReleases is sending notifications on new releases.