Skip to main content

CEEMS Exporter

Flags

Flag / Environment VariableDescriptionDefault
--helpShow context-sensitive help (also try --help-long and --help-man).
--versionShow application version.
--log.formatOutput format of log messages. One of: [logfmt, json]logfmt
--log.levelOnly log messages with the given severity or above. One of: [debug, info, warn, error]info
--web.systemd-socketUse systemd socket activation listeners instead of port listeners (Linux only).false
--web.debug-serverEnable /debug/pprof profiling endpointsfalse
--runtime.gomaxprocs / GOMAXPROCSThe target number of CPUs Go will run on1
--collector.disable-defaultsSet all collectors to disabled by defaultfalse
--web.max-requestsMaximum number of parallel scrape requests. Use 0 to disable.40
--web.disable-exporter-metricsExclude metrics about the exporter itself (promhttp_, process_, go_*).false
--web.targets-pathPath under which to expose Grafana Alloy targets./alloy-targets
--web.telemetry-pathPath under which to expose metrics./metrics
--web.config.file / CEEMS_EXPORTER_WEB_CONFIG_FILEPath to configuration file that can enable TLS or authentication. See: docs.md
--web.listen-addressAddresses on which to expose metrics and web interface.:9010
--collector.slurmEnable the slurm collectorfalse
--collector.redfishEnable the redfish collectorfalse
--collector.raplEnable the rapl collectortrue
--collector.meminfoEnable the meminfo collectortrue
--collector.libvirtEnable the libvirt collectorfalse
--collector.ipmi_dcmiEnable the IPMI DCMI collectorfalse
--collector.emissionsEnable the emissions collectorfalse
--collector.cray_pm_countersEnable the Cray PMC collectorfalse
--collector.cpuEnable the cpu collectortrue
--collector.slurm.gpu-order-mapGPU order mapping between SLURM and NVIDIA SMI/ROCm SMI tools. It should be of format <slurm_gpu_index>:<nvidia_or_rocm_smi_index>[.<mig_gpu_instance_id>] delimited by ",".
--collector.slurm.psi-metricsEnables collection of PSI metricsfalse
--collector.slurm.swap-memory-metricsEnables collection of swap memory metricsfalse
--collector.redfish.web-config / CEEMS_EXPORTER_REDFISH_COLL_CONFIG_FILEPath to Redfish web configuration file.
--collector.rdma.statsEnables collection of RDMA statsfalse
--collector.rapl.enable-zone-labelEnables RAPL zone labelsfalse
--collector.perf.env-varEnable profiling only on the processes having any of these environment variables set. If empty, all processes will be profiled.
--collector.perf.cache-profilersperf cache profilers to collect
--collector.perf.hardware-cache-eventsEnables collection of perf hardware cache eventsfalse
--collector.perf.software-profilersperf software profilers to collect
--collector.perf.software-eventsEnables collection of perf software eventsfalse
--collector.perf.hardware-profilersperf hardware profilers to collect
--collector.perf.hardware-eventsEnables collection of perf hardware eventsfalse
--collector.meminfo.all-statsEnable collecting all meminfo statsfalse
--collector.libvirt.psi-metricsEnables collection of PSI metricsfalse
--collector.libvirt.blkio-metricsEnables collection of block IO metricsfalse
--collector.libvirt.swap-memory-metricsEnables collection of swap memory metricsfalse
--collector.ipmi_dcmi.force-native-modeForce native mode using OpenIPMI driver.false
--collector.ipmi_dcmi.dev-numDevice number used by OpenIPMI driver. For e.g. if device is found at /dev/ipmi0, device number is 00
--collector.ipmi_dcmi.cmdIPMI DCMI command to get system power statistics. Use full path to executables.
--collector.emissions.providerExports emission factors from these providers (default: all). Supported providers: - "owid": Our World In Data - "emaps": Electricity Maps - "rte": RTE eCO2 Mix (Only for France)
--collector.ebpf.fs-mount-pointFile system mount points to monitor IO stats. If empty all mount points are monitored. It is strongly advised to choose appropriate mount points to reduce cardinality.
--collector.ebpf.network-metricsEnables collection of network metrics using ebpffalse
--collector.ebpf.io-metricsEnables collection of IO metrics using ebpffalse
--collector.cgroup.active-subsystemActive cgroup subsystem for cgroups v1.cpuacct
--discoverer.alloy-targets.self-profilerEnable continuous profiling by Grafana Alloy on current processfalse
--discoverer.alloy-targets.env-varEnable continuous profiling by Grafana Alloy only on the processes having any of these environment variables
--discoverer.alloy-targetsEnable Grafana Alloy targets discovererfalse