-
Bug
-
Resolution: Done
-
Critical
-
None
-
False
-
None
-
False
-
NEW
-
VERIFIED
-
Before this update, pod metadata could be altered by fluent plugins because the values passed through the pipeline by reference. This update fixes that to ensure each log message receives a copy of the pod metadata so each message processes independently.
-
- Install RHOL 5.5
- Configure by default the CLO (not needed to configure the CLF)
- Deploy to a new application
$ oc new-project ocasalsa-test $ oc new-app rails-postgresql-example
- Check in kibana that for the pods deployed, the most of the times, the `kubernetes.flat_labels` are not populated and other they are for the same pod without doing any changes
Install RHOL 5.5 Configure by default the CLO (not needed to configure the CLF) Deploy to a new application $ oc new -project ocasalsa-test $ oc new -app rails-postgresql-example Check in kibana that for the pods deployed, the most of the times, the `kubernetes.flat_labels` are not populated and other they are for the same pod without doing any changes
-
Log Collection - Sprint 224, Log Collection - Sprint 225
-
Moderate
In RHOL 5.5, the `kubernetes.flat_labels` are not always populated.
I should expect that the `kubernetes.flat_labels` should be always present like it was in RHOL 5.4, in all the logs and not sometimes yes and sometimes not.
NOTE:
The collector pods are not giving any error related to the `kubernetes.flat_labels`
Then only error visible in the collector pods is the same that described in https://issues.redhat.com/browse/LOG-2859 .
- links to
- mentioned on
(1 mentioned on)
