- Fix issue preventing host process info from being reported.
0.0.31 / 2017-06-19
- Configure alternate proc path
- Configure alternate etc path
- Fix logging of "Next Metadata Send" message
0.0.30 / 2017-03-25
- Fix memory leak related to cpu utilization metrics collection
- Upgrade to latest signalfx-python v1.0.16
0.0.29 / 2016-11-02
- Added python dependency six
0.0.28 / 2016-10-18
- Upgrade to latest signalfx-python v1.0.7
- Log outdated metrics instead of emitting them
0.0.27 / 2016-09-21
- Periodically check if the plugin is running on AWS
0.0.26 / 2016-09-08
- Add optional configuration to report cpu utilization on a per core basis
0.0.25 / 2016-08-03
- Upgrade to latest signalfx-python v1.0.4
0.0.24 / 2016-06-28
- no longer report gauge.sf.host-dpm as it was an instantaneous rate of change and comparisons with it were inaccurate.
0.0.23 / 2016-06-02
- upgrade to latest signalfx-python
0.0.21 / 2016-04-04
- send version information as dimensions on the plugin_uptime metric
- work around issue where collectd segfaults on shutdown when write callback is an instance method
- add way to turn off datapoints
- add dither to initial notifications
0.0.20 / 2016-03-31
- change memory.utilization to not include cached, bufferred or slab but just be used/total
- fix repetative logging when types.db.plugin isn't found
0.0.19 / 2016-03-25
- add ability for multiple endpoints
0.0.18 / 2016-03-21
- refactory utilizations
- add unit tests
0.0.17 / 2016-03-17
- quantize utilization metrics for better accuracy
0.0.16 / 2016-03-15
- fix bug where memory.utilization wasn't always reported
0.0.15 / 2016-03-11
- fix bug with summation metrics first points being wild
- detect intervals for each type of utilization data and emit on same schedule
- add AWSUniqueId if not present, and we're an AWS box
0.0.14 / 2016-03-09
- fix bug with utilization metrics not showing without DPM
- fix bug where plugin doesn't emit metrics on old collectds
0.0.13 / 2016-02-29
- add Utilization metrics behind flag defaulted to true
- update sf.host-response.max to include the timeout times
0.0.12 / 2016-01-15
- change uptime to plugin uptime
- quiet a few more top exceptions
- default timeout to 3 seconds
0.0.11 / 2016-01-07
- fix units issue on memory
- add metrics for max response time, send errors and dpm
- quiet top info when pids disappear
- steal hosts from other notifications even if we aren't the one sending them
0.0.10 / 2015-11-04
- add dogstatsd support if configured to do so
0.0.9 / 2015-10-26
- use self instead of os.getpid to get collectd version
- change how often we send metadata from only on startup to startup, one minute, hour and day, then per day from then on
- use psutil instead of top
- send in plugin version as plugin_instance
0.0.8 / 2015-10-02
- wait one interval before sending notifications
- url error parsing more resilient
- add back in logical cpus
- better linux version checking
0.0.7 / 2015-09-29
- Remove dependence on requests
0.0.6 / 2015-09-27
- Add interfaces notifcation
- Remove logical cpus
- prefix with host_
- support /etc/os-release for linux_version
- Send LargeNotifs if message too large for collectd
0.0.5 / 2015-09-25
- Use host from collectd for Top Info notification
- Add mem_total for physical memory in kb
0.0.4 / 2015-09-20
- Top on amazon linux and better error handling
0.0.3 / 2015-09-20
- Changed format of top-info payload to include version
- Added ability to turn off process table collection
0.0.2 / 2015-08-25
- Removed proof-of-life datapoint's plugin instance and changed name to host-uptime
- Made metadata plugic work from cmdline
- Added linux_version and collectd_version to meta-data
- Added Top Data sent
0.0.1 / 2015-08-18
- Send notifications when configured to
- Collect cpu, system and ec2 information and emit as notifications
- Emit single metric for proof of life