The RHDH OCP Orchestrator Helm nightly e2e jobs (e2e-ocp-helm-nightly) were failing due to two issues in .ci/pipelines/utils.sh:

1. Missing SSL support for SonataFlow database creation — The helm chart's create-sonataflow-database job does not include the PGSSLMODE environment variable, causing database creation to fail when connecting to external PostgreSQL instances that require SSL (Crunchy Data PostgreSQL). A manual database creation workaround was added that runs a psql pod with SSL support and proper /tmp write access.
2. Hardcoded PostgreSQL namespace in RBAC helm values — The externalDBHost in values_showcase-rbac.yaml was hardcoded to postgress-external-db-primary.postgress-external-db.svc.cluster.local, but nightly Prow jobs create the postgres namespace as postgress-external-db-nightly. This caused DNS resolution failures and CrashLoopBackOff for sonataflow-platform-data-index-service and sonataflow-platform-jobs-service. Fixed by passing a --set override in perform_helm_install to inject the correct hostname from the NAME_SPACE_POSTGRES_DB environment variable.

Additional cleanup: removed redundant PGSSLMODE export and verify_sonataflow_database function, added timeout to the database job wait loop.

PR: https://github.com/redhat-developer/rhdh/pull/3929