30m is too long and there is a risk of running out of disk space/inodes completely if something is filling up disk very fast (like log file).
Signed-off-by: Vitaly Zhuravlev <v-zhuravlev@users.noreply.github.com>
* Migrate from Python codespell to golangci-lint misspell.
* Inline errcheck exclude list in the golangci-lint config.
Signed-off-by: Ben Kochie <superq@gmail.com>
This change adds the ability to process multiple stat calls in parallel.
Processing is rate-limited based on the new flag
`collector.filesystem.stat-workers` (default 4).
Caveat: filesystem stats information is no longer in the same order as
returned by `/proc/1/mounts`. This should not be an issue.
Caveat: This change currently uses unbuffered channels to prove
correctness without reliance on buffers. Buffered channels will yield
superior performance.
Signed-off-by: Erica Mays <erica@emays.dev>
* [CHANGE] Fix cpustat when some cpus are offline #2318
* [CHANGE] Remove metrics of offline CPUs in CPU collector #2605
* [CHANGE] Deprecate ntp collector #2603
* [CHANGE] Remove bcache `cache_readaheads_totals` metrics #2583
* [CHANGE] Deprecate supervisord collector #2685
* [FEATURE] Enable uname collector on NetBSD #2559
* [FEATURE] NetBSD support for the meminfo collector #2570
* [FEATURE] NetBSD support for CPU collector #2626
* [FEATURE] Add FreeBSD collector for netisr subsystem #2668
* [FEATURE] Add softirqs collector #2669
* [ENHANCEMENT] Add suspended as a `node_zfs_zpool_state` #2449
* [ENHANCEMENT] Add administrative state of Linux network interfaces #2515
* [ENHANCEMENT] Log current value of GOMAXPROCS #2537
* [ENHANCEMENT] Add profiler options for perf collector #2542
* [ENHANCEMENT] Allow root path as metrics path #2590
* [ENHANCEMENT] Add cpu frequency governor metrics #2569
* [ENHANCEMENT] Add new landing page #2622
* [ENHANCEMENT] Reduce privileges needed for btrfs device stats #2634
* [ENHANCEMENT] Add ZFS `memory_available_bytes` #2687
* [ENHANCEMENT] Use `SCSI_IDENT_SERIAL` as serial in diskstats #2612
* [ENHANCEMENT] Read missing from netlink netclass attributes from sysfs #2669
* [BUGFIX] perf: fixes for automatically detecting the correct tracefs mountpoints #2553
* [BUGFIX] Fix `thermal_zone` collector noise @2554
* [BUGFIX] Fix a problem fetching the user wire count on FreeBSD 2584
* [BUGFIX] interrupts: Fix fields on linux aarch64 #2631
* [BUGFIX] Remove metrics of offline CPUs in CPU collector #2605
* [BUGFIX] Fix OpenBSD filesystem collector string parsing #2637
* [BUGFIX] Fix bad reporting of `node_cpu_seconds_total` in OpenBSD #2663
Signed-off-by: Ben Kochie <superq@gmail.com>
Read missing dev_id, name_assign_type, and addr_assign_type
from sysfs, since they only take a device-specific lock and
not the whole RTNL lock. This means reading them is much less
impactful on other system processes than many of the other
attributes in sysfs that do take the RTNL lock.
Signed-off-by: Dan Williams <dcbw@redhat.com>
On most hard drives, `ID_SERIAL_SHORT` and `SCSI_IDENT_SERIAL` are identical,
but on some SAS drives they do differ. In that case, `SCSI_IDENT_SERIAL`
corresponds to the serial number printed on the drive label, and to the value
returned by `smartctl -i`.
So use that value by default for the `serial` label on the `node_disk_info`
metric, and fallback to `ID_SERIAL_SHORT` only if it's undefined.
Signed-off-by: Benoît Knecht <bknecht@protonmail.ch>
Mark the `supervisord` as deprecated. This process
supevisor, like `runit`, is of scope for the node_exporter.
Signed-off-by: Ben Kochie <superq@gmail.com>
* bcache: remove cache_readaheads_totals metrics #2103
Signed-off-by: Saleh Sal <0xack13@gmail.com>
* Append bcacheReadaheadMetrics when CacheReadaheads value exists
Signed-off-by: Saleh Sal <0xack13@gmail.com>
* Update test cases for cachereadahead greater than zero
Signed-off-by: Saleh Sal <0xack13@gmail.com>
---------
Signed-off-by: Saleh Sal <0xack13@gmail.com>
Use the filesystem collector for all OpenBSD archs, there is no reason to
only use it on amd64 systems.
Signed-off-by: Claudio Jeker <claudio@openbsd.org>
This updates the annotation for the NodeClockSkewDetected mixin alert to
match the new threshold set.
Original discussion was in this PR: https://github.com/prometheus/node_exporter/pull/1480
I spent an embarrassingly large amount of time trying to figure out how
the heck that alert would mean 300s of clock skew. Turns out the
annotation was just left the same after the threshold change.
Signed-off-by: Will Bollock <wbollock@linode.com>
Vitaly Zhuravlev