Releases: DataDog/lading
Releases · DataDog/lading
v0.17.2-rc5
Changed
- We now clamp CPU utilization to number of cores, not scaled to.
v0.17.2-rc4
Changed
- We now elide sampling threads that are not also processes in the observer.
v0.17.2-rc3
Changed
- Correct
rss
to be reported in terms of bytes, not pages, as before.
v0.17.2-rc2
Removed
- Observer no longer emits tick data for kernel and user-space time.
Changed
- Observer now distinguishes between parent and children processes.
- Config can now be specified using an env var
LADING_CONFIG
. If set, the env var takes precedence over the on-disk config file.
v0.17.2-rc1
Changed
- We now expose two crates from this project, one for capture payloads and the
other the main lading project.
v0.17.1
Fixed
- Removed a leftover debug print in the Json payload implementation.
v0.17.0
Changed
- Adjusted the default throttle to stable from predictive
- On Linux calculate CPU utilization in terms of logical, not physical, cores when possible.
- CPU percentage calculated in the same manner as Agent's
- Throttle metrics are now labeled with the respective generator's labels.
- Observer now calculates CPU utilization with respect to target cgroup hard, soft limits.
v0.17.0-rc3
Changed
- On Linux calculate CPU utilization in terms of logical, not physical, cores when possible.
v0.17.0-rc2
Changed
- CPU percentage calculated in the same manner as Agent's
v0.17.0-rc1
Changed
- Throttle metrics are now labeled with the respective generator's labels.
- Observer now calculates CPU utilization with respect to target cgroup hard, soft limits.