💡 Enhancements 💡
clusterReceiver
: For the option,clusterReceiver.eventsEnabled
, the logs pipeline for k8s_events now adds attributes of the typek8s.<objectkind>.name
andk8s.<objectkind>.uid
. (#1691)
For example, if the log k8s event is about object typeStatefulSet
, the exported log to Splunk will have these 2 additional attributes:
The existing attributesk8s.statefulset.name: value(k8s.object.name) k8s.statefulset.uid: value(k8s.object.uid)
k8s.object.kind
,k8s.object.name
andk8s.object.uid
are still present.
In addition to these, if the event is for kind Pod, and the k8s.object.fieldPath has a specific container spec, the log will have an additional attributek8s.container.name
with the value of the container name.
🧰 Bug fixes 🧰
agent
: Do not setup the entities pipeline if Splunk Observability isn't enabled. (#1699)agent
: a fix for a scenario where some logs might be missed due to the pod log file being rolled over during high load, set featureGates.fixMissedLogsDuringLogRotation to true to enable the fix (#1690)all
: Restore values ofservice.name
resource attribute for internal metrics changed in 0.120.0 (#1692)
The value ofservice.name
resource attribute was changed tootelcol
due to a library upgrade
in the Prometheus receiver. This change restores the values that were set before the based on the
collector mode:otel-agent
,otel-collector
orotel-k8s-cluster-receiver
.