prometheus/node_exporter 0.16.0 / 2018-05-15



  • Breaking changes

    This release contains major breaking changes to metric names. Many metrics have new names, labels, and label values in order to conform to current naming conventions.

    • Linux node_cpu metrics now break out guest values into separate metrics. See Issue #737
    • Many counter metrics have been renamed to include _total.
    • Many metrics have been renamed/modified to include base units, for example node_cpu is now node_cpu_seconds_total.

    In order to help with the transition we have an upgrade guide.

    Other breaking changes:

    • The megacli collector has been removed, is now replaced by the storcli.py textfile helper.

    • The gmond collector has been removed.

    • The textfile collector will now treat timestamps as errors.

    • [CHANGE] Split out guest cpu metrics on Linux. #744

    • [CHANGE] Exclude Linux proc from filesystem type regexp #774

    • [CHANGE] Ignore more virtual filesystems #775

    • [CHANGE] Remove obsolete megacli collector. #798

    • [CHANGE] Ignore /var/lib/docker by default. #814

    • [CHANGE] Cleanup NFS metrics #834

    • [CHANGE] Only report core throttles per core, not per cpu #836

    • [CHANGE] Treat custom textfile metric timestamps as errors #769

    • [CHANGE] Use lowercase cpu label name in interrupts #849

    • [CHANGE] Enable bonding collector by default. #872

    • [CHANGE] Greatly reduce the metrics vmstat returns by default. #874

    • [CHANGE] Greatly trim what netstat collector exposes by default #876

    • [CHANGE] Drop exec_ prefix and move node_boot_time_seconds from exec to new boottime collector and enable for Darwin/Dragonfly/FreeBSD/NetBSD/OpenBSD. #839, #901

    • [CHANGE] Remove depreated gmond collector #852

    • [CHANGE] align Darwin disk stat names with Linux #930

    • [FEATURE] Add collect[] parameter #699

    • [FEATURE] Add text collector conversion for ipmitool output. #746

    • [FEATURE] Add openbsd meminfo #724

    • [FEATURE] Add systemd summary metrics #765

    • [FEATURE] Add OpenBSD CPU collector #805

    • [FEATURE] Add NFS Server metrics collector. #803

    • [FEATURE] add sample directory size exporter #789

    • [ENHANCEMENT] added Wear_Leveling_Count attribute to smartmon.sh script #707

    • [ENHANCEMENT] Simplify Utsname string conversion #716

    • [ENHANCEMENT] apt.sh: handle multiple origins in apt-get output #757

    • [ENHANCEMENT] Export systemd timers last trigger seconds. #807

    • [ENHANCEMENT] updates for zfsonlinux 0.7.5 #779

    • [BUGFIX] Fix smartmon.sh textfile script #700

    • [BUGFIX] netdev: Change valueType to CounterValue #749

    • [BUGFIX] textfile: fix duplicate metrics error #738

    • [BUGFIX] Fix panic by updating github.com/ema/qdisc dependency #778

    • [BUGFIX] Use uint64 in the ZFS collector #714

    • [BUGFIX] multiply page size after float64 coercion to avoid signed integer overflow #780

    • [BUGFIX] smartmon: Escape double quotes in device model family #772

    • [BUGFIX] Fix log level regression in #533 #815

    • [BUGFIX] Correct the ClocksPerSec scaling factor on Darwin #846

    • [BUGFIX] Count core throttles per core and per package #871

    • [BUGFIX] Fix netdev collector for linux #890 #910

    • [BUGFIX] Fix memory corruption when number of filesystems > 16 on FreeBSD #900

    • [BUGFIX] Fix parsing of interface aliases in netdev linux #904

    A retrouver ici : https://github.com/prometheus/node_exporter/releases/tag/v0.16.0

    Post généré automatiquement



  • Bon, cette mise à jour impose un peu de boulot 🙂 Je viens de faire la mise à jour et si vous utilisez, par ex, les dashboard Grafana node-exporter qu’on trouve sur grafana.com, il y a, comme annoncé, pas mal de métriques qui changent (cpu, memory, disk et network).
    Bref, il faudra mettre à jour à la main ou re-importer les dashboard.
    Idem pour l’alerting Prometheus, il faut revoir les règles !



  • Update : on trouve pas mal de dashboard compatibles 0.16 sur grafana.com !