sustainable-computing-io/kepler
Kepler (Kubernetes-based Efficient Power Level Exporter) uses eBPF to probe performance counters and other system stats, use ML models to estimate workload energy consumption based on these stats, and exports them as Prometheus metrics
GoApache-2.0
Issues
- 1
Prometheus deployed with high granularity scrape interval not set correctly
#1750 opened by jaylenwang7 - 0
- 8
GPU Collector Does Not Use DCGM nor NVML
#1783 opened by marvin-steinke - 1
Mock ACPI CI doesn't checkout the PR code base
#1709 opened by vprashar2929 - 2
kepler exporter pod stuck at ContainerCreating
#1784 opened by smaitra53 - 2
Error Loading eBPF Objects (softirq_entry)
#1765 opened by marvin-steinke - 2
Update DynPower Model to 0.7.11
#1748 opened by sthaha - 2
Kepler latest fails to create Power Model to estimate Node Platform/Component Power on VM
#1663 opened by vprashar2929 - 1
Update the data weight to point to latest models
#1680 opened by vprashar2929 - 2
add the custom `metal-runner-action@custom-action` in sustainable-computing org
#1739 opened by vimalk78 - 0
Kepler reports abnormally high platform power on Equinix
#1774 opened by sthaha - 0
Kepler does not report metrics on resources outside of some system namespaces
#1771 opened by BoyanBanev - 0
Add check for building images on PR
#1759 opened by vprashar2929 - 0
- 0
validator produces wrong MAPE
#1688 opened by sthaha - 1
Update Models and Model Server to 0.7.11 for Compose
#1711 opened by KaiyiLiu1234 - 0
- 0
Add README for mock-acpi workflow
#1713 opened by vprashar2929 - 0
- 10
integration_test / integration_test_with_libbpf blocks PR
#1669 opened by sthaha - 5
kepler-action v0.0.8 does not deploy clusters
#1697 opened by dave-tucker - 6
bpf sampling no longer reduces overhead
#1607 opened by rootfs - 2
- 0
v0.7.11 x86 w/PCP: kepler_node_platform_joules_total readings very low in Perf CoPilot
#1674 opened by jharriga - 0
Add the possibility to use metadata in libvirt to customize the vm_id of VM metrics
#1625 opened by aharivel - 1
- 1
- 2
- 6
High Kepler CPU usage under normal workloads
#1670 opened by vimalk78 - 1
Reduce eBPF Probe Execution Time
#1611 opened by dave-tucker - 0
- 6
- 1
Proposal: Go-based Model Server
#1657 opened by dave-tucker - 0
validator: upgrade paramiko when CryptographyDeprecationWarning: TripleDES is fixed
#1667 opened by sthaha - 0
- 0
Kepler latest underreporting metric values
#1641 opened by vprashar2929 - 0
- 0
validator report MAPE as inf
#1633 opened by sthaha - 1
reporting UNKNOWN cpu_architecture - kepler v0.7.11 on RHEL9/ARM64 Ampere system
#1616 opened by jharriga - 0
Metric kepler_node_package_joules_total on RHEL9/aarch64 only reports mode=dynamic, not "idle"
#1617 opened by jharriga - 0
- 0
validator results interpretation
#1602 opened by rootfs - 7
- 0
- 0
- 1
Kepler not running in Raspberry Pis
#1597 opened by ravaga - 0
- 0
Remove overflow checks in Uint64Stat
#1593 opened by dave-tucker - 0
Add kepler_node metrics comparison for BM/VM
#1582 opened by KaiyiLiu1234 - 1
BPF related metrics were zero
#1580 opened by YaSuenag