Fixed Issues

Lists the issues that were fixed in the Cloudera Observability release.

OBS-12466: GPU utilization and memory data now available for Cloudera AI workloads on OpenShift Container Platform: Previously, when using Cloudera AI workloads on OpenShift Container Platform clusters, GPU utilization and GPU memory consumption data were not displayed in the Cloudera Observability UI. This occurred because the metrics collector used the host's network namespace, and the default OpenShift Container Platform network policy prevented the local Prometheus from receiving GPU metrics from pods using the host network.; This issue is now resolved by updating the network policy to allow ingress from host-network pods to non-host-network pods.
OBS-12802: Metrics collector pods fail to start after upgrading Cloudera Observability: This issue is now resolved in Cloudera Data Services on premises 1.5.5 SP4.
OBS-5212: Spark jobs are not shown on the Jobs & Queries tab of the Cloudera Observability Real-time monitoring user interface: The issue occurs when Cloudera Data Hub clusters are created with versions from 7.2.18.100 to 7.2.18.400. The issue has been fixed on Cloudera Data Hub version 7.2.18.500 or higher.