
Ephemeral test databases for each PR. With HDI containers and instance templates, I can spin a clean HANA schema per pipeline, load fixtures, run integration tests, and tear it down. Ridiculously fast set-based data setup. The in-memory column store + SQLScript lets me generate millions of realistic rows (orders, line items, events) in minutes. Great for perf/regression runs without waiting on upstream systems.
Smart Data Access for “virtual” dependencies. I can federate into S/4HANA sandboxes or flat files and keep tests end-to-end without copying everything. It cuts flakiness from stale replicas.
Observation that actually helps. HANA Cockpit + expensive statements trace + plan visualizations make it obvious which queries are hurting our E2E timings. Quick wins: compression settings, column store indexes, and parameterized plans.
Security that doesn’t block testing. Roles, analytic privileges, and schema isolation let us mimic production access patterns in tests (RBAC, row-level filters) so auth bugs surface early.
Elasticity when we need it. Scale up for load tests, scale down for regular CI. Data tiering to the data lake keeps costs sane for big historical fixtures. Review collected by and hosted on G2.com.
Privilege model can be a maze. HDI container roles + classic schema privileges are powerful but confusing; onboarding new QA engineers takes time.
SQLScript quirks. Great for set logic, less fun for procedural control flow and unit testing small routines.
Tooling sprawl. Cockpit, BAS, CLI, DP Agent… it works, but context-switching is real compared to a single Postgres+psql workflow. Review collected by and hosted on G2.com.