1173636 - stop using the collectd aggregation module

Status: RESOLVED FIXED

(Infrastructure & Operations :: RelOps: Puppet, task)

Description

[Jake Watkins [:dividehex]]

Cpu metrics aggregation can be done in the cpu module itself now.  This will workaround the 'double up' bug we experienced with the aggregation module.

https://github.com/collectd/collectd/issues/297

Comment 1

[Jake Watkins [:dividehex]]

Attachment: bug1173636-1.patch

Configuration change to collectd cpu modules; sums up all cpu cores and cpu active states into a single percentage.  This is what was being attempted via the aggregation module and doesn't require the graphite derivative func to be applied to convert jiffies to percentage.

This will also start a new metric name for all hosts under cpu.percent.active

Comment 2

[Jake Watkins [:dividehex]]

Comment on attachment 8622804 [details] [diff] [review]
bug1173636-1.patch

Also adds cpu module back to osx since it was busted in previous versions with a temperature read value error.  Once this lands we can remove the associated papertrail filter

Comment 3

[Jake Watkins [:dividehex]]

Comment on attachment 8622804 [details] [diff] [review]
bug1173636-1.patch

remote:   https://hg.mozilla.org/build/puppet/rev/3e3358cf38c3
remote:   https://hg.mozilla.org/build/puppet/rev/eebc9068348f