What's Changed

✨ Added

feat(devices): Added device interfaces by @vimalk78 in https://github.com/sustainable-computing-io/kepler/pull/1960
feat: add initial skeleton by @vprashar2929 in https://github.com/sustainable-computing-io/kepler/pull/1961
feat(version): add version information to build by @sthaha in https://github.com/sustainable-computing-io/kepler/pull/1965
feat(logger): use slog for logging by @sthaha in https://github.com/sustainable-computing-io/kepler/pull/1964
feat(config): config and arg parsing using kingpin by @sthaha in https://github.com/sustainable-computing-io/kepler/pull/1970
feat(ci): add pre-commit hooks by @vprashar2929 in https://github.com/sustainable-computing-io/kepler/pull/1973
feat: add Prometheus exporter and API server to Kepler by @sthaha in https://github.com/sustainable-computing-io/kepler/pull/1986

🐛 Fixed

fix(server): escape endpoint description by @vprashar2929 in https://github.com/sustainable-computing-io/kepler/pull/1987

Full Changelog: https://github.com/sustainable-computing-io/kepler/compare/v0.0.0-reboot...v0.0.1-reboot

Consumption - Computation and Communication - Go
Published by github-actions[bot] 21 days ago

Kepler - release-0.8.0

b7add5dc fix: false detection of Grace Hopper GPU
e3fb5298 fix(openshift): patch deployment issue
b43260a9 fix(collector): data race when updating maps simultaenously
92b85921 feat(config): Make /proc and /sys paths configurable
7212f9af fix: Redfish: Use Chassis instead of Systems to obtain chassis ID
e86161e3 Fix: Bump golang.org/x/net from 0.28.0 to 0.34.0
1440cc88 feat: add TLS endpoint to kepler exporter
98275d04 [fix]: enhance kubelet package impl (#1894)
c9c2aba2 Fixes paths for grace arm
c9524a83 [fix]: try to reduce escapes to heap (#1886)
cf469c26 feat(sensor): support NVIDIA Grace Hopper
662b0d52 fix: pre-commit auto update commit author (#1873)
a447077f [fix]: adjust gosec permissions settings according to GHA ref (#1867)
e45f6c45 [fix]: add test case with darwin OS (#1856)
2ddabe27 [fix]: fix with permissions for PR pipeline
133f219a [fix]: update gha permission settings
59b2cfae [fix]: fix image libbpf build version (#1860)
16fdd9f2 [fix]: base image elfutils version issue (#1859)
46272c6a fix(cpu): fixed reading cpus.yaml
11711e79 feat(config): allow config dir to be passed as argument (#1850)
9ef9706a fix(yamllint): remove excess spaces
e2c3c6f8 fix(yamllint): disable yamllint quote rule
f4f5eb46 clean(process-exporter): Add yamllint fixes
fdcea6b3 fix(yamllint): Fix yamllint errors
35e160b2 feat(process-exporter): Add process-exporter to dev and metal
75b95332 fix(bpf): exclude swapper process bpf_cpu_time (#1830)
6a968cd3 fix: epel-release installation via rpm
2beddafb feat: move from ubi to ubi-minimal
fe78fd55 feat(validator): incorporate MAE in validatons
91c8e627 fix: remove .dockerignore to ensure clean version computation (#1809)
26de449f feat(metrics): refactor idle power exposure

New Contributors

@maichouni-mitek made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1825
@github-actions made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1875
@witekest made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1907
@vishnuchalla made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1943

Full Changelog: https://github.com/sustainable-computing-io/kepler/compare/v0.7.12...v0.8.0

Consumption - Computation and Communication - Go
Published by github-actions[bot] about 1 month ago

908db28d fix(bpf_collector): fix command name in case of kernel processes
b90a63b7 fix(bpf): exclude bpf overhead in bpf_cpu_time
987a1391 fix(metrics): Remove resource usage check for skipping bpf metrics
c20c23ce fix: error initializing dcgm
41213764 fix(aa66ada): reading libraries to the builder
4c63b433 fix(validator): update the bpf cpu time query
097541c4 fix: gosec failures (#1778)
5b95acb5 fix: nvml/dcgm builds
9d14581a fix: habana image build
aa66ada3 fix(dockerfile): remove redundant habanalabs installation steps
34d27b81 fix: do not probe for power-meters when disabled
ecd5f545 feat(models): update acpi dyn to 0.7.11
7303454d feat(models): update acpi abspower to 0.7.11
fcc8e0a1 feat(models): update intel-rapl abspower to 0.7.11
c84bfd0b feat(models): log model source url
7b077624 feat: get trainer from model_name in weight
a7b9892b fix(models): add predictor name in errors
aa822eef feat: compute core ratio for local regressor (#1743)
5d6ecc05 fix(config): trim spaces and new lines in MODEL_CONFIG
b2926d1b fix: limit max core ratio to 1
28d42a4f feat: compute idle power with core ratio (#1732)
d8a6c14f feat: add machine spec generator/reader for model weight request
11ff51d1 feat: add --disable-power-meter option
414533ca fix: set default trainer only for local regressor
d4462319 fix: format ComponentModelWeights
a6f75a45 feat: add model_name attribute to ComponentModelWeights
6858c58c fix: watcher resubmit items to workqueue (#1686)
eb5a72a9 fix(bpf): Fix overhead when sampling (#1685)
a97030f9 fix(bpf): use prev_tgid to register process
5432a394 feat: customize vm_id with libvirt metadata
73367d3c fix: correct regex path name for VM
6a6017b4 feat: save validation result as json and show the static dashboard using js
c99e3993 fix(pkg/bpf): Use channel to process events (#1671)
cde78331 fix: resolve pid 0 to system_processes
b4246074 feat(kubernetes): Use workqueues
a39ae555 fix: typo in filename
d73094ed fix: apply suggestions from code review
5a113b2a fix(bpf): tgid is in the upper 32 bits

New Contributors

@arthurus-rex made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1621

Full Changelog: https://github.com/sustainable-computing-io/kepler/compare/v0.7.11...v0.7.12

Consumption - Computation and Communication - Go
Published by github-actions[bot] 7 months ago

Kepler - release-0.7.11

Changes

150669106d956 - feat(validator): trigger validator workflow on changes (#1591)
c7b3ddb40088c - fix(collector): convert cpu time in collection time instead of reporting time to avoid inconsistent use of cpu time in models
9c80387af3f91 - bpf: account all running state processes (#1546)
91fc8d4fd5511 - feat(validator): Add workflow for validator tests (#1570)
a2289d24355f3 - fix: fallback to reading cpus.yaml relative to current dir (#1572)
bfaadae9d7e3e - pick up the go mod vendor changes
fb7ef358a98f4 - feat(metrics): selectively expose prom metrics to reduce overhead
d412bfb4d5041 - fix: vendor/github.com/jaypipes/ghw/Dockerfile to reduce vulnerabilities (#1578)
365ac035f7ade - bpf: remove tgid map
c427a477e0705 - fix(manifest): uncomment openshift SCC (#1575)
ec2a775ca3381 - fix(validator): improve the validator config sample (#1569)
0e22839e53d1f - fix: update the VERSION variable assignment method (#1552)
96dd443f99d71 - fix: Fix uncomment of YAML in hack/build-manifests.sh
8931d616a4772 - feat(validator): load validations from validations.yaml
4a7bc31509580 - fix(compose): enable bpf cgroup id
a57041c39ff6d - fix(bpf): Fix kepler_write_page_cache attach
fbe9b3cabbba1 - fix(bpf): Access __state from task_struct (#1550)
0b0b215818cbe - fix(bpf): Use BTF-Defined Raw Tracepoints (#1542)
a08a5f6d26e1d - deps: Fix usage of textparse.NewPromParser
aad6964bd979c - fix(bpf): Fix map lookup for IRQ/Page Cache
9114e75f9ae86 - Fix MSE and MAPE Single Queries (#1522)
330a5315bc382 - fix(bpf): restore command label in process metrics
edd4d04649144 - review feedback: fix mse queries
d6420d5dbd658 - bump up local_dev_cluster_version version
4ced508166639 - bpf-collector: change log verbosity to easily show it in CI
34889bb4f8fe5 - libbpf: update to use microseconds instead of milliseconds in the ebpf code because the low precision is identifying that the precess was not active
4337a5eedbffa - bpf: remove task time
0426e8f7646e1 - feat(exporter): Graceful Shutdown
aec3ab5a40b98 - report validator results
07636b196dba3 - Replace expected and actual query with single query (#1489)
1759cca4cdbc9 - feat(compose): add build arguments for Kepler image
c678217dd2ae8 - use pmu name to get arm cpu id since archspec does not help here
8bb405f90c1ff - stats: update the verbosity of annoying key error message due to missing gpu metrics (#1480)
59af568a85b0a - Add Test Cases for Prometheus, Config, Stresser for Validator (#1461)
bdd44b15edf42 - bpf: fix the process parameter order to match the c and go code (#1479)
747e7eb5b1fb4 - fix: ensure all entries from bpf map is copied (#1477)
c092204a3d191 - make: quote ldflags
468ed25a61e9a - add vm name option to validator (#1474)
244ae8bd7da94 - feat: expose version label in kepler_build_info (#1473)
6ae21a06fb58d - update validator usage; remove job from prom query
3ac4f6b309845 - feat(cgroup): Add podman support (#1455)
ada7884b916f4 - fix platform power return unit (#1468)
3c7e777bd5074 - fix(collector): Fix use of waitgroups
b134a849ec5f2 - fix(cmd/validator): Don't add when passing a wg
0158b0bec23c8 - fix(dev-dashboard): update and correct metrics in dev dashboard
f92532ac0077c - add new maintainers per 05/21 community meeting vote results (#1462)
efad46fba69a1 - provide a simple template for maintainer nominate (#1463)
9e957f3c1a0a4 - finish kepler on rhel tests
dcf78e67a83f9 - fix: remove logging while collecting GPU metrics
49acca9a5d259 - fix(model): Use correct variable in IsNodeComponentPowerModelEnabled() (#1458)
1b93eb1d8a4a7 - Adding New Metric Cases to Case module (#1453)
ea3e2f80b0ae0 - add equinix metal instance to CI
53d06d49d3afa - add PR review bot (#1446)
0baec479033aa - feat: Fixed eBPF Feature Detection (#1443)
5f59172c569df - fix(bpf): cleanup initialising structs and nested ifs (#1444)
2bca8dc61f005 - update hack/libbpf-headers.sh script to pull v1.3.0

New Contributors

@wenboown made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1451
@caniszczyk made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1578

Full Changelog: https://github.com/sustainable-computing-io/kepler/compare/v0.7.10...v0.7.11

Consumption - Computation and Communication - Go
Published by github-actions[bot] 10 months ago

Kepler - release-0.7.10

Summary

fix(bpfassets): Fix object file lookup (#1419)
feat(bpf): Build for bpfel and bpfeb
feat(bpf): Bump up libbpf to 1.3.0
fix(dashboard): show metal and VM metrics correctly (#1395)
doc(dev): add section on how to profile (#1396)
feat(bpf): Portable eBPF Probes
feat(test): initial version of validator tool
dev(compose): add manifests for validation
fix(collector): Fix Segmentation fault when collecting CPU Freq from BPF (#1387)
feat(kepler): enable pprof (#1383)
fix habana installation
fix previous pid of finish_task_switch (#1370)
fix: update dashboard for docker-compose
fix(build): reduce image size by squashing install and clean steps
feat(compose): add docker-compose for easier local development
feat(exporter): log listening port
fix(build): reduce container image size (#1336)

New Contributors

@dave-tucker made their first contribution in https://github.com/sustainable-computing-io/kepler/pull/1384

Full Changelog: https://github.com/sustainable-computing-io/kepler/compare/v0.7.9...v0.7.10

Consumption - Computation and Communication - Go
Published by github-actions[bot] 12 months ago