Answer
It helps to think of the category in layers rather than a single feature. The replication engine, usually journal-based on IBM i, is the foundation, moving changes from production to the secondary system as they occur. On top of that sits monitoring, which tells the team whether the secondary copy can actually be trusted at any given moment, since replication that silently falls behind or drops objects is arguably worse than no replication at all because it creates false confidence. Above that sits the failover procedure itself, the actual mechanics of promoting the secondary system to production, and finally the testing discipline that proves the whole stack works before the business has to rely on it during a real crisis.
Buyers evaluating vendors should ask about each layer separately rather than accepting a single "yes we do HA" answer, because products vary widely in how much of this stack they deliver versus how much the customer has to assemble themselves through custom scripting or manual procedure documents. Ask specifically what a role swap looks like in practice, how the vendor's monitoring would have caught their own worst historical outage, and how often their reference customers actually run failover tests versus simply trusting the replication is healthy. The gap between vendors shows up far more in these operational details than in the base replication technology, which has become fairly commoditized across the market.