The Splunk OTel Collector ingests infrastructure metrics via dedicated receivers for Linux hosts, Kubernetes clusters, and Prometheus targets.
The hostmetrics receiver collects OS-level CPU, memory, disk, filesystem, load, and network metrics directly from Linux /proc and sysfs. In Kubernetes, kubeletstats pulls pod and container resource metrics from the local node Kubelet API, while k8s_cluster gathers cluster-wide state (deployments, replica sets, node conditions) from the Kubernetes API server.
For applications already instrumented with Prometheus client libraries, the prometheus receiver scrapes /metrics endpoints on configured scrape_configs intervals, translating Prometheus gauges and counters into OTLP metrics without rewriting application code. All collected metrics are commonly exported to Splunk Infrastructure Monitoring via the signalfx exporter.
receivers:
hostmetrics:
collection_interval: 10s
scrapers:
cpu:
memory:
disk:
prometheus:
config:
scrape_configs:
- job_name: "app-metrics"
scrape_interval: 10s
static_configs:
- targets: ["localhost:9090"]