We want to align the service_name logic of the K8s Monitoring Helm Chart Pod Logs feature on OpenTelemetry resource attributes specs on Kubernetes.

For this, we need to retrieve the deployment/cronjob/job of the pods through the "controller-by" attribute.
If the OpenTelemetry Collector K8s Processor is capable of retrieving these metadata, it's not clear if Alloy's discovery.relabel can also do it.

• https://github.com/grafana/k8s-monitoring-helm/blob/v2.0.28/charts/k8s-monitoring/charts/feature-pod-logs/templates/_common_pod_discovery.alloy.tpl
• The source code of the logic to retrieve pod metadata seems to be https://github.com/prometheus/prometheus/blob/main/discovery/kubernetes/pod.go

OpenTelemetry logic for service.name

Choose the first value found:

• pod.annotation[resource.opentelemetry.io/service.name]
• pod.label[app.kubernetes.io/instance] (well-known label app.kubernetes.io/instance)
• pod.label[app.kubernetes.io/name] (well-known label app.kubernetes.io/name)
• k8s.deployment.name
• k8s.replicaset.name
• k8s.statefulset.name
• k8s.daemonset.name
• k8s.cronjob.name
• k8s.job.name
• k8s.pod.name
• k8s.container.name

See also

• Revert logs service_name preference of container.name vs k8s.pod.name #1533