v3.0.8

Changed

• Ensure all telegraf metrics are timestamped in microseconds. Because output to
  telegraf is buffered, we can't rely on telegraf doing the timestamping
  (multiple samples get timestamped with the same or very similar times), so we
  add a timestamp to every metric line as soon as we see it.
• ntpmon_info metric in telegraf mode split into
  ntpmon_resident_set_size_bytes, ntpmon_virtual_memory_size_bytes, and
  ntpmon_uptime_seconds in prometheus mode, with tags only on
  ntpmon_uptime_seconds, to tidy metric names and reduce tag cardinality when
  used with prometheus. These retain their existing names in telegraf mode.
• Make messages from log file tailer tidier and more consistent.
• Correctly flag /etc/default/ntpmon as a configuration file in the Debian
  package.

Full Changelog: v3.0.7...v3.0.8

v3.0.7

• Handle loss of telegraf socket more gracefully
• Handle failure of subprocess more gracefully
• Ensure that systemd always restarts ntpmon regardless of exit code

Full Changelog: v3.0.6...v3.0.7

v3.0.6

Added

• ntpmon_info metric in prometheus and telegraf modes, including tags for
  various system, python, and ntp components.
  • If someone wants this for collectd and can explain how to do it in a way
    which makes sense, please get in touch.
• --version command line argument.
• Roadmap in README.
• Versioning strategy in CHANGELOG.

Changed

• Fix data type on stratum metric for ntpd. This was an integer under 2.x and
  now is consistently so between chronyd and ntpd.
• Use peertype instead of type for individual peer metrics, to provide tag
  compatibility between ntpmon_peer and ntpmon_peers metrics.
• Fix python 3.8 compatibility with debug flag.

Full Changelog: v3.0.5...v3.0.6

v3.0.5

What's Changed

• 3.0.5 release by @paulgear in #36

Full Changelog: v3.0.0...v3.0.5

v3.0.0

https://www.libertysys.com.au/2023/12/an-update-on-ntpmon/

What's Changed

• Added individual peer stats (derived from chronyd's measurements.log or ntpd's peerstats) to metrics output. These are automatically detected in the default locations of /var/log/chrony/measurements.log and /var/log/ntpstats/peerstats and emitted automatically to telegraf or the prometheus exporter.
• Added --logfile command line option to allow changing the above defaults.
• Reformatted all code with 'black' (see Makefile 'format' target). Please use this Makefile target before submitting any pull requests.
• Updated license to AGPLv3.
• Removed trace-related metrics.
• Removed juju layer.
• Clean up internal structure:
  • outputs.py now provides a class structure for encapsulating knowledge about collectd, prometheus, and telegraf
  • alert.py and ntpmon.py no longer have special casing for the mode
  • all argument processing and defaults selection is done in get_args()
• Some collectd data types have been updated to reflect the defaults available in version 5.12. I'm considering deprecating collectd support in a near future version. If you are using ntpmon with collectd and would prefer this support not to go away, please let me know.

Full change set: v2.1.0...v3.0.0

v2.1.0

ChangeLog

• Read the explanation of the changes at https://github.com/paulgear/ntpmon/blob/main/CHANGELOG.md

What's Changed

• Add prometheus listener fixes by @paulgear in #22
• Fix version specification in dependency by @paulgear in #23
• v2.0.2+repack1 by @paulgear in #24
• Rebuild by @paulgear in #25
• Don't send metrics with no value in nagios perfdata by @flyingflo in #26
• Prepare 2.1.0 release by @paulgear in #27

New Contributors

• @flyingflo made their first contribution in #26

Full Changelog: v2.0.1...v2.1.0

v2.0.2

Update changelog