- [CHANGE] Remove broken metrics collection #68
- [CHANGE] Remove use of cgroups #105
- [FEATURE] Add unit timestamp metrics #58
Now released under the Prometheus Community
- [CHANGE] systemd_service_restart_total changed label name from
type
toname
#32 - [CHANGE] Remove prommod metrics #36
- [CHANGE] Rename collector flags #53
- [FEATURE] Add TLS and Basic Auth #54
- New option
--collector.user
, which allows users to connect to the user systemd instance. Thanks @pelov :) - New feature
--collector.enable-ip-accounting
, enables service ip accounting metrics. This feature only works with systemd 235 and above. thanks @jwhited :)
systemd_unit_state
labeltype
has new meaning Now shows Unit type (service
,scope
, etc), not Service Unit types (simple
,forking
, etc) or mount unit types(aufs
,ext3
, etc). Service and mount types have been moved tosystemd_unit_info
- [CHANGE] Important! Docker images are now published to
povilasv/systemd-exporter
andquay.io/povilasv/systemd-exporter
registry. - [FEATURE] Add support for amd64, 386, arm, arm64, mips, mipsle, mips64, mips64le, ppc64, ppc64le, s390x docker images.
- [FEATURE] Read unit CPU usage from cgroup. Added
systemd_unit_cpu_seconds_total
metric. Note - Untested on unified hierarchy - [FEATURE] Add
systemd_unit_info
with metainformation about units incl. subtype specific info - [ENHANCEMENT] Added
type
label to all metrics namedsystemd_unit-*
to support PromQL grouping
- [ENHANCEMENT]
systemd_unit_state
works for all unit types, not just service and mount units - [ENHANCEMENT] Scrapes are approx 80% faster. If needed, set GOMAXPROCS to limit max concurrency
- [CHANGE] Start tracking metric cardinality in readme
- [CHANGE] Expanded default set of unit types monitored. Only device unit types are not enabled by default
- [BUGFIX]
timer_last_trigger_seconds
metric is now exported as expected for all timers
- [CLEANUP] Introduced changelog. From now on, changes will be reported in this file.