Releases: google/cadvisor
v0.38.0
- #1594 - chore: add storage_driver_buffer_duration in Influxdb storage docs
- #1924 - add hugepages info to attributes
- #2578 - Add perf event grouping.
- #2590 - Use current Docker registry
- #2611 - Aggregate perf metrics
- #2612 - Add stats to stdout storage
- #2618 - Update to containerd v1.4.0-beta.2 and runc v1.0.0-rc91
- #2621 - Memory numa stats
- #2627 - use Google Charts loader and not jsapi
- #2631 - Add entry for libpfm related tests to Makefile
- #2632 - Handling zeros in readPerfStat
- #2638 - Add stats to statsd storage
- #2639 - Add logs and simplify setup of raw perf events
- #2640 - Remove exclude guest flag from perf event attrs.
- #2644 - Use perf attributes from unix lib.
- #2646 - Fixed https proxy issue by installing 'full' wget in Docker alpine-based build stage
- #2655 - Update readme to point to discuss.kubernetes.io
- #2659 - Fix ordering of processes table
- #2665 - add clean operation when watchForNewContainers/Start failed
- #2669 - Update release documentation and process
- #2676 - Fix runtime error when there are no NVM devices.
- #2678 - Add checking checksum of libpfm4
- #2679 - Fix typo in libipmctl
- #2682 - Add missing flag to runtime_options.md
- #2683 - Add flags that were not previously published
- #2687 - Move mount library dependency from utils/mount to mount-utils
- #2689 - Increase the readability of perf event logs.
- #2690 - Try to read from sysfs before giving up on non-x86_64
- #2691 - Broken build configuration when custom build tags are used
- #2695 - Add information about limits of opened perf event files.
- #2697 - Update to new docker(v19.03.13) and containerd(1.4.1)
- #2702 - Increase golang ci lint timeout to 5 minutes
- #2706 - Add a badge for the current e2e test result
- #2707 - Fix Avoid random values in unix.PerfEventAttr{}
- #2711 - validateMemoryAccounting: fix for cgroup v2
- #2713 - Bump golang to 1.15
- #2714 - update docker client method
- #2716 - Update dependencies
Docker Image: N/A
gcr.io Image: gcr.io/cadvisor/cadvisor:v0.38.0
SHA256: 53a02d669ae20bfb9634639c1315d4823c06f6cafe15e625973abedce6c99084
v0.35.1
v0.37.0
- Add on-demand collection for prometheus metrics
- Fix detection of image filesystem
- Fix disk metrics for devicemapper devices
- Add NVM Power and NVM, Dimm, memory information to machine info
- Fix detection of OOM Kills on 5.0 linux kernels
- Add support for perf core and uncore event monitoring
- Add hugetlb container metrics
- Split into multiple go modules
- Add referenced memory metrics
- Publish images to gcr.io/cadvisor instead of gcr.io/google_containers
- Add socket id to numa topology in machine info
- Add resource control (Resctlr) metrics
Docker Image: N/A
gcr.io Image: gcr.io/cadvisor/cadvisor:v0.37.0
SHA256: 4f4aeee957e64c5ccccefb12c6b4dec283fb307fa420c864ebabfcfd9d6ce051
v0.36.0
- Add support for risc and mips CPUs
- Add advanced TCP stats
- Fix bug in which cAdvisor could fail to discover docker's root directory
- The stdout storage driver now supports metric timestamps
- Add ulimit metrics
- Support multi-arch container builds
- Switch to go modules
Docker Image: N/A
gcr.io Image: gcr.io/google_containers/cadvisor:v0.36.0
SHA256: 7bbc3e229ea41940efe25c37a7c0f6010a4d7e47bee98e0f8e0359a4690b39be
v0.35.0
- Add hugepage info per-numa-node
- Add support for cgoups v2 unified higherarchy
- Drop support for rkt
- Fix a bug that prevented running with multiple tmpfs mounts
Docker Image: N/A
gcr.io Image: gcr.io/google_containers/cadvisor:v0.35.0
SHA256: e9bb5513d9c49741a6ccf9ec0ce68a0ca5d25d8e75cb426c525b003fbf19cdd9
v0.34.0
- Fix disk stats in LXD using ZFS storage pool
- Support monitoring non-k8s containerd namespaces
- The
storage_driverflag now supports comma-separated inputs - Add
container_sockets,container_threads, andcontainer_threads_maxmetrics - Fix CRI-O missing network metris bug
- Add
disable_root_cgroup_statsflag to allow not collecting stats from the root cgroup.
Docker Image: N/A
gcr.io Image: gcr.io/google_containers/cadvisor:v0.34.0
SHA256: 060c6361dd6d4478ff0572e8496522d8189cf956eea2656b6247ad683abcc9d3
v0.33.0
- Add --raw_cgroup_prefix_whitelist flag to allow configuring which raw cgroup trees cAdvisor monitors
- Replace
duandfindwith a golang implementation - Periodically update MachineInfo to support hot-add/remove
- Add explicit timestamps to prometheus metrics to fix rate calculations
- Add --url_base_prefix flag to provide better support for reverse proxies
- Add --white_listed_container_labels flag to allow specifying the container labels added as prometheus labels
Docker Image: google/cadvisor:v0.33.0
gcr.io Image: gcr.io/google_containers/cadvisor:v0.33.0
SHA256: 098441fad760b2b66c4109fd942d0bda815e7b52de0a7470cd0e087440727892
v0.32.0
- Add container process and file descriptor metrics (disabled by default)
- Rename
typelabel tofailure_typefor prometheusmemory_failures_totalmetric - Reduce mesos error logging when mesos not present
Docker Image: google/cadvisor:v0.32.0
gcr.io Image: gcr.io/google_containers/cadvisor:v0.32.0
SHA256: 62419c0e06edb55a9c02e68fcae3a81abac2a2d98122c36a9124259e0ca8916c
v0.31.0
- Fix NVML initialization race condition
- Fix brtfs filesystem discovery
- Fix race condition with AllDockerContainers
- Don't watch .mount cgroups
- Reduce lock contention during list containers
- Don't produce prometheus metrics for ignored metrics
- Add option to not export container labels as prometheus labels
- Docs: Publish cAdvisor daemonset
- Docs: Add documentation for exported prometheus metrics
Docker Image: google/cadvisor:v0.31.0
gcr.io Image: gcr.io/google_containers/cadvisor:v0.31.0
SHA256: 195faa8d3953b4534bb0f050738ca3f7d6b1a646a1663e5660c9848967ff678f
v0.30.2
- Fix concurrent map access panic
- Don't watch .mount cgroups to reduce number of inotify watches
- Fix NVML initialization race condition
- Fix brtfs disk metrics when using a subdirectory of a subvolume
Docker Image: google/cadvisor:v0.30.2
gcr.io Image: gcr.io/google_containers/cadvisor:v0.30.2
SHA256: a5fc548b39edf2fb90bc887e15b38ba9d171f869ff55b858647373db86bce766