This repository has been archived by the owner on Oct 23, 2024. It is now read-only.
v4.18.3
- Make the new
load
monitor send non-cpu scaled values by default. There is now a flagperCPU
that can be enabled (default false) if you want load scaled to the number of CPUs on the system. This makes the newload
monitor match the values of the oldcollectd/load
monitor. - Fix the name of the
disk-io
monitor in the default configs - Add the
disk_op.total
metric to thedisk-io
monitor that was sent by thecollectd/signalfx-metadata
monitor - Make duration config values safer by interpreting plain integers as seconds instead of nanoseconds, which was almost certainly wrong if a user had previously specified a duration as such.
Breaking Changes:
- If you have switched to the
load
monitor, the values will no longer be scaled per cpu by default. This was considered a bug.
Docker Image:
quay.io/signalfx/signalfx-agent:4.18.3
(digest:sha256:be9c44bc406dbec15c013cdc5edffec750c400f74d34e7c13b80afbb669fce3c
)