Fixed issues in 1.5.5 SP4

Fixed issues for Cloudera Data Engineering on premises 1.5.5 Service Pack 4 resolve bugs and functional discrepancies previously logged through Cloudera Support or identified as Known Issues by customers and Cloudera Quality Engineering teams.

DEX-18287: Git repository synchronization failed due to expired TLS certificates

Previously, Cloudera Data Engineering did not support updating TLS certificates for existing Git repositories. If a repository's TLS certificate expired or was rotated, the Git repository synchronization failed, requiring users to delete and recreate the repository. This issue is now resolved by adding support to replace PEM certificate bundles for existing Git repositories through the Cloudera Data Engineering UI, CLI, or Runtime API without recreating the repository.

DEX-22761: YuniKorn gang scheduling failed due to missing native sidecar resource calculations

Previously, Fluent Bit native sidecar resources were omitted from YuniKorn gang scheduling placeholder calculations for Spark drivers and executors. Because native sidecars contributed to a pod's total resource requests, the actual driver and executor pods exceeded their reserved placeholder sizes. This discrepancy caused YuniKorn to cancel the placeholders and fail gang scheduling. This issue is resolved by including Fluent Bit native sidecar CPU and memory requests in placeholder resource calculations.

DEX-23419: Cloudera Data Engineering backup restore failed for TLS-enabled Git-repo resources

Previously, restoring a Cloudera Data Engineering backup failed for Git-repo resources that had TLS certificates enabled. This occurred because the backup process stored a reference to a cluster-local Kubernetes secret instead of the actual certificate data, causing a base64 encoding error during the restore process on a different virtual cluster. This issue is now resolved by inlining the certificate data as base64-encoded text directly into the backup archive, ensuring successful restoration across clusters.